This materiaW is provided by the lnternational Center for Living Aquatic " j

Resources Manageaeta (ICLARM) an indepenotent, nonprofit research c,, i, tsr working on fisheries nd aquaculture in tropical third-orld countries.

A­. c p,.o. boo ot, alkati, metro manila, philippines[\Ii ICLARN/ STUD IES A

Research and Education for the Development of Integrated Crop-Livestock-Fish Farming

Systems in the Tropics

P. Edwards R.S.V. Puilin J.A. Gartner

k~

4-. -. . 1-...'..

r,

'~-,

MINTER N TIO .CA CR F R L V A. E,

INTERNATIONAL CENTER FOR LIVING AQUATIC RESOURCES MANAGEMENT

Research and Education for the Development of Integrated Crop-Livestock-Fish Farming

Systems in the Tropics

P. Edwards R.S.V. Pullin J.A. Gartner

1988

INTERNATIONAL CENTER FOR LIVING AQUATIC RESOURCES MANAGEMENT MANLA, PHIUPPINES

1988

Research and education for the development of integratedcrop-livestock-fish farming systTm~sin the tropics

P. EDWARDS R.S.V. PuIIIN J.A. GARTNER

Published by the International Center for Living Aqual'c Resources Management, MC P.O. Box 1501, Makati, Metro Manila, Philipptnes with financial assisLnce from the United Nations i)evelopment F-ogramrre New York, USA

Printed in Manih, Philippines

Edwards, P., R.S.V. Pullin and J.A. Gartnmr. 1988. Research and education for the develt'ment of integrated crop-livestock-fishfarming ystems in the tropics. ICLARM Studies and Reviews 16, 53 p. intemational Center for Living Aquatic Resources Management, Manila, Philippines.

ISSN 0115-4389 ISBN 971-1022-46-X

Cover: Small-scale integrated crop-livestock-fish farming in a rainfed at-a ofNortheast Thailand. This rice farm has a small fishpondthat provides fish, permits dry season cultivation of vegetable, on the dikes and supplies drinking water for livestock.

ICLARM Contribution No. 470

Contents

Preface..................................................

Introduction.1...............................................

The Concept of Integrated Farming Systems

A Definition of Integrated Farming ............................... 5 The Development of Integrated Farming Systems. .................... 5 Classification of Farming Systems ................................... 7 CharaLteristics of Asian Farming Systems .............................. 9

Research 'ramcwork

General Considerations ................................... 11 Crop Subsystems ...................................... 11 Livestock Subsystem.. .................................. 12 Fish Subsystems ........................................ 13

Strategies for increasing production ............................ 13 Extensive, semi-intensive and intensive aquaculture .............. .... 13 Interactions in crop-iivestock-fish integrated farming systems ........ ... 15 Deriving fish culture practices from natural aquatic ecosystems ........ .. 16 Choice of fish species ....................................... 17 Fishpond dynamics ........................................ 21 Types of nutritiona! inputs to ponds ............................ 23 Physical characteristics of the pond ............................. 26 Pond sediments ................................... 27 Fish stock management ................................ 28 Systems modelling .................................. 29 Potential health hazards ............................... 30

Education

A Systems Approach to Agricultural Development ........................ 31 Education Programs ..................................... 32 The Need for Research in Association with Tertiary Education ............ ... 34 An Example of a Systems Approach to Education in Integrated Farming ...... .. 34

Background and relation to national programs and other institutions. . .. 34 The M.Sc. program in agricultural systems at AIT ............... .... 35

Generalprinciples................................... 35 Structure................................... 35Importantareasof concern.............................. 36 Some probiems............................... 37

Employment Opportunities ........................................ 38 Points of Vulnerability .................................... 38

iii

An Institutional Framework

General Considerations ................................... 40 Geographical/climatic ................................ 40 A cademic . ....................................... 41 Institutional .43 On-farm activities .................................. 45

An Institutional Framework ..................................... 46

Acknowledgements .......................................... 48

Referenct.s ............................................... 49

Appendix I: Workshop - Towards a Research Framework for Tropical Integrated Agriculture-Aquaculture Farming Systems, 15-17 October 1986, Manila, Philippines ...... 52

iv

Preface Successful integration of aquaculture with agriculture is a complex subject, not leastbecause of the poorly developed research base for aquaculture in comparison with that of

agriculture. Aquaculture science is a relatively new field of study; moreover, it is clear from theresearch output and teaching programs of most institutions involved in the subject that, despite aprofession of interest in integrated agriculure-aquaculture, the attention of their scientistsremains narrowly focused on the fish and the aquatic environment ratler than on the farmer andthe whole farm. This is not s,-prisinrg as most of the professional scientists involved in aquaculture were educated as life scientists (principally zoologists) or as specialists in aquaticscience (particularly fisheries) and ihey prefer to stay within their primary disciplines.

An atempt is made in this publication to create a framework for a truly interdisciplinaryaipproach to research and education in integrated fanning - a fusion c,,"agricultural andaquaculture sciences. It has been prepared by two aquatic biologists, albeit experienced inintegrated fanning research Ond education, and an agriculturist with a special interest in farmingsystems. I recognize that an aquatic bias still shows through in this document, but I believe that itis unioue as a first step towards a formal integration of the sciences supporting integratedagriculture-aquaculture fanning systems.

Integration of aquaculture with agriculture is more developed in Asia than in any otherregion of the world. However such integrated fanning systems are presently used by only a verysmall minority of farmers (<I%) in a few countries and have not progressed far in terms ofproductivity and efficiency from their traditional beginnings. This point is o','ten missed bydonors and development agencies who, seeing that Asia produces 75% of the world's culturedfish production and that Chinese integrated fanning principles have a long and successfulhistory, fail to recognize that vast potential still exists for many more of Asia's numerous andneedy small-scale fanners to enjoy the benefits of integration of aquaculture into farmingsystems. To realize this potential requires a new research and education program, as is proposed _n this publication.

For Africa, the potential for aquactulture and integrated farming development is far lesscertain. In the growing campaign by donors and research and development agencies to dcvelopappropriate systems to improve the nutrition and livelihood of African peoples, there appears tobe a tacit assumption that aquaculture and integrated farming systems incorporating aquaculturehave high potential for development in Africa. This assumption has little basis in fact. For manyAfrican nations there are serious constraints to aquaculture and integrated fanning development,such as adverse environmental conditions for fish growth (aridity, high altitude/lowtemperature); underdeveloped/shifting agriculture; labor shortages; lack of interest in fishhusbandry; competition from capture Fsheries in fish markets; socia! attitudev/t0boos and other factors.

It is, of course, probable that integrated farming systems incorporating fish will flourish in some African countries in which major constraints are absent or surmountable and for which the necessary resea:'ch and production trials can be undertaken. Meanwhile, a cautious approach toaquaculture development is needed; not a rush into development by transfer of foreigntechnologies. Such a cautious approach should best be undertaken in parallel with furtherresearch for tile development of Asian integrated crop-fish and crop-livestock-fish systems forwhi 7h reliable management guidelines are still lacking.

What are the prospects that this new approach will succeed? I believe that they are excellent,given adequate recognition by scientists and donors of the potential of integrated farming and the

V

need for research and education to underpin its development. The United Nations DevelopmentProgramme already has taken an admirable lead by supporting ICLARM to produce this studyand engage in related planning work. Other donors are beginning to support integrated farmingresearch and development, including aquaculture components, in Africa as well as in Asia. On a most encouraging note, the Consultative Group on International Agricultural Research (CGIAR)and its Technical Advisory Committee (TAC) have begun to take a strong interest in aquacultureand its potential for responding to more substantial research. A CGIAR involvement in aquaculture research would greatly increase contact between agricultural and aquact lture scientists and thus benefit integrated farming programs. The professional expertise that would beneeded to pursue this research is becoming available as more economists and social scientists trained in agricultural institutions are now taking an interest in the integration of aquaculturewith agriculture.

Finally, one might ask why combine Research and Education? It is my view that the two are inseparable if the develement of integrated farming systems is to succeed and thereby make a greater contribution to food supply and livelihood. The agricultural and aquaculture researchers needed for the work outlined here must educate themselves, and a new generation of professionals committed to this broader view of integrated farming systems must merge.

I would like to take this opportunity to thank many colleagues around the world who helpedreview early drafts of this manuscript. Your ideas and reactions have helped the authors considerably.

IAN R. SMITH Director General, ICLARM July 1988

vi

Research and Education for the Development of Integrated

Crop-Livestock-Fish Farming Systems in the Tropics

P. EDWARDS Professorof Aquacuhure

Asian Institute of Technology GPO Box 2 754

Bangkok 10501, Thailand

R.S.V. PULLIN Director,Aquaculture Program

InternationalCenterfor Living AquaticResources Management

MC P.O. Box 1501, Makati Metro Manila, Philippines

J.A. GARTNER Associar Professorof FarmingSystems

Asian Institueof Technology GPOBox 2 754

Bangkok 10501, Thailand

Introduction There is a pressing need to increase the productivity and profitability of farming indeveloping countries. According to the study "Agriculture Towards 2000" (FAO 1981), in whichthe fiture development of world agriculture was analyzed, by the year 2000 the world populationwill be more than 6 billion (it passed the 5 billion mark in July 1987) and the demand foragricultural products in third-world countries will double. Small-scale farmers comprise the bulkof the population of the developing world and the challenge is to raise their productivity per unitarea per unit of time and per unit of capital input because the amount of arable land per person isdeclining due to overpopulation. Productivities are also declining due to environmental

I

2

degradation; for example, deforestation, leading to more unstable water resources and sol erosion. There is a need to improve the efficiency of utilization of the limited resource base of small-scale farmers through the promotion of integrated farming to improve their diet, balance risks among various farming subsystems, provide fuller employment and generate surplusproduce for sale. Indeed integrated farming systems offer great prospects for the development of sustainable third-world agriculture with minimal adverse environmental impact. This is particularly relevant for those farms which are operated under rainfed conditions. These comprise about 70-80% of the total agricultural land.

Grigg (1974) studied the evolution of the major fanning systems of the world as he believed that a knowledge of the past would enable us to better understand present faning systems and aid their future development. In a second major treatise, he considered the relationship between population growth and agrarian change from a historical perspective (Grigg 1980). He stressed that the need for successful production responses in agriculture in the developing world is imperative. It is ironical that the food production potentials of the tropics, where year-round production is often possible, have not often been realized. Although most third-world farmers grow monocultures of rice or maize, there is a wide variety of traditional cropping patterns in the humid tropics. The production potentials of the latter remain to be fully exploited (Hoque 1984). Hoque referred to crops but the same applies even more to livestock and fish, the other subsystems in an integrated crop-livestock-fish farming system.

Why include a fish subsystem in an integrated farm? Fish have many advantages as farm produce. They are a highly nutritious and valuable traditional food in much of Asia and Africa. Fish provide about 25% of animal protein for human consumption in Africa and from 28 to 80% in South and Southeast Asian countries. Fish are smal!, valuable units of saleable protein and once grown can be kept alive on maintenance diets with little or no loss of condition for later harvesting at will. They are excellent converters of low grade feeds into high quality animal protein because, unlike terrestrial livestock, they do not need to use dietary energy to maintain body temperature or posture.

A holistic farming systems approach is taken in this study because the greatest potential for fish culture lies with farmers who are already engaged in the production of crops and livestock. The idea is to bring aquaculture to resource-poor, small-scale farmers who have limited access to the off-farm inputs necessary to exploit modern farming technology. Fish are produced byrecycling byproducts of agronomy and anima! husbandry into animal protein. Nutrient-rich pond water and mud are potential resources for adjacent crop products. Aquaculture thereby becomes the third partner alongside existing crop and livestock farming subsystems on small-scale farms. The cost of raising fish in such integrated farming systems would be lower than iii systems using pond inputs from agro-industry and would be feasible for small-scale farmers.

The grea test sccpe for the development of integrated crop-livestock-fish farming systems is in the humid tropics. This is where the need is also greatest. Trewartha's modification of the Kippen System of classification of climates (Money 1978) is used in this review (Fig. 1). Tropical climates are defined as those with the mean temperature of the coolest month greater than 180C (Money 1978: Oldeman and Frbre 1982). This allows tropical fish to grow year­round. Three types of tropical climate are recognized in the system:

1. A Rainy Climate with either no dry season or a short dry season, but with enough rainfall to support rainforest (and therefore fodder) year-round;

2. A Wet and Dry Climate (tropical savannah) with distinct wet and.dry seasons; and 3. A Semi-Arid Climate with a long dry season and a short rainy season. Aquaculture is possible in climates I and 2 but generally not in climate 3. Water shortages

in climates 2 and 3 could constrain not only fish culture but also the availability of pond inputs (forage and byproducts from crop and livestock subsystems).

The integrated farming systems discussed in this study make use of tropical fish, particularly the omnivorous tilapia which has been hailed as the "aquatic chicken" of the future (Pullin1985a). Tilapias breed ane grow year-round in the tropics. The Subtropic Humid Climate, with year-round rainfall and the mean temperature of the coldest month between 0 and 180C is included in Fig. I because it covers China, the origin and a current exponent of integrated crop­

3

Tropical climates

wet and dry

semi-arid Subtropical climate

g~

0 1,500 km 0

Sce

le'lrewamlaFig. 1. "I' I 1 1odilicationof [tic K6ppen sysic i I i,, ,,fIIl:r ta ih ad ,taa Ihurelacauconcepllt ifredu ion ,,fall clevations tosea level. \Vithin Illese clilajitic /)11'N, i h

pINnIlal '.;itemIa, .jltt k­. r Lrfy livtiKa ..fsh f'amIIs cpenad unaIlaw(l c'ujj(aafi s (adlapled frroI Money I197S).

livestock-fish farming. However, only eurythermal warmwater fish species can normally becultured in China. The growing season varies from 8 to 11 months in the Yangts. and PearlRiver Basins, respectively, because of seasonally low winter temperatures. Research for thedevelopment of tropical integrated farming systems should be conducted in the tropicsunhindered as far as possible by seasonal climatic constraints.

Recognizing then the scope for development of integrated crop-livestock-fish farmingsystems in the tropics, what information and expertise are available? Both are very limited;hence ihe need for a program of research and education as outlined in this study. Tropicalaquaculture has a weak research base (Pullin and Nea! 1984) compared with agronomy andanimal husbandry. This weakness has been exacerbated in integrated farming research anddevelopment by the narrowly-based educational background and hence narrow ,ision of manyresearchers and developers. For example, most aquaculture scientists understand and see only thefish and their requirements and do not take into account the complete resource system.A holistic view of the farm is essential. Aquaculturists must learn to understand existingcrop and crop-livestock farming sysems and agricultural researchers the fish farming subsystem.The processes of research and education for the development of integrated farming systems aretherefore closely interlinked. This calls for an innovative approach to bring aquaculture into themainstream of agriculture. There is a need for researchers and educators who, while seeking the

4

inputs of specialist crop, livestock and aquatic scientists to answer specific questions, are themselves generalists,researching and educating others by means of a whole farm perspective and a broad interdisciplinary mix of biology, economics and social science.

There is a vast literature on crop and livestock production in tropical third-world countries but comparatively little on integration of these with aquaculture. Information "n crop-livestock­fish integrated farming systems has been collected by Temprosa et al. (in prep.) as an annotated bibliography. Published information and the expertise of the participants of a workshopconvened in 1986 to set the scene for a new program of research and education for the development of crop-livestock-fish integrated farming systems were used for this study(Appendix I). The evaluation and development of farming systems, the identification of the target groups, geographical scope, and potential impact of the incorporation of a fish subsystem are considered first. Detailed frameworks are then given for research and education, followed by a discussion on institutional aspects.

The Concept of Integrated Farming Systems A Definition of Integrated FarmA'"

The word integrated is de the Latin verb "integrare" which means to make whole,to complete by addition of parts, u: ,v combine parts into a whole. The crop, livestock and fishsubsystems may function independently in certain farming systems, and their products be onlyadditive. However, an output from one subsystem in an integrated farming system whichotherwise may have been wasted becomes an input to another subsystem resulting in a greaterefficiency of output of desired products from thc land/water area under a farmer's control. Thereis synergism in integrated faning since the working together of the subsystems has a greater tora effect tnan the sum of their individual effects.

The main bio!ogical feature of an integrated farming system is byproduct recycling; butimproved space utiiliation, in which two subsystems occupy part or all of the space required for one subsystem, may be an important aspect of increased productivity. A major socioeconomic benefit of integrated farming is that inputs to the various subsystems that comprise the farmingsystem tend to be intra-farm, with a diminished reliance on inter-farm or agro-industrial inputs.Integrated farming systems also spread the risks associated with farming because of theincreased diversity of produce. They a!o lead to a more balanced diet for the farming family that chooses to eat some of its own produce.

The Development of ntegratedFarmingSystems

A schema is presented in this study of the possible evolutionary development of integratedfarming systems to set the research framework recommended here in an appropriate context (Fig.2). The three major ca.,egories of farming - settled agriculture, shifting cultivation and pastoralnomadism - are adopted from an example of a classification of world farming systems bySpedding (i979). However, settlerJ agriculture is divided here into three phases - cropdominated, integrated crop/livestock and industrial monoculture - to emphasize the role thatintegrated farming systems can play in bringing aquaculture to resource-poor, small-scale farmers with limited access to often costly off-farm inputs. The rationale for the three phases ofsettled agriculture is derived largely from studies of the agricultural farming systems of theworld and their evolution by Whittlesey (1936), Duckham (1959, 1966), Duckham andMasefield (1971) and Grigg (1974, 1980). A simplified schema cannot represent all possiblevariants of the world's agricultural systems (Spedding 1979) but it is a useful conceptualframework for thi, study.

Hunting/gathering/fishing preceded the development of agriculture but are still ofimportance in many third-world countries, particularly with regard to fish. Indeed, the capture ofwild fish, as opposed to aquaculture, is still the major source of fish in most third-world countries.

Shifting cultivation involves periodic shifts t-) new land as the fertility of the original patchis exhausted. This is now confined main':, ,o mountainous areas. There is little potential forintegration with aquaculture because of the restricted area suitable for pond construction inmountain terrain and because of the migrations of the society. The earliest form of agriculture inthe tropics was thought to be "vegeculture", based on vegetative propagation of roots (taro,cassava, yams, sweet-potatoes, and arrowroot) and the collection of tree fruits such as bananas

5

Hunting/Gathering/ Fishing

Shifting Cultivation

Settled Agriculture 1 Pastoral (Crop Dominated) Nomadism

Settled Agriculture 2 1 (Integrated Crop/Livestock)

I-I Settled Agriculture 3

(Industrial Monoculture)

Fig. 2.A schema of possible evolutionary phases in farming systems devclojxnent.

and coconuts. Pigs and poultry were probably domesticated as scavengers. Vegeculture was displaced after about 3,500 BC by "seed" agriculture, based or, wet-rice cultivation, with largeruminants as draught animals; it survives only in remote areas and in mixed gardens (Grigg 1974).

In densely populated pre-industrial societ-ies, most land is under food crops, :!nd livestock are kept mainly as draught animals; scavenging pigs and poultry may be kept as .source of meat. This farming system is called Settled Agriculture Phase 1 (Crop-Dominated), Most of western European agriculture was in this phase until about 1850. The main crop was cereals and fertility of land was maintained by the two and later the three field system in which one strip was left fallow each year to rest the land. There was little or no integration between subsystemsbecause cattle were kept mainly for draught purposes and depended on rough grazing. Most third-world countries are still in Settled Agriculture Phase 1, except for those areas affected bythe "Green Revolution" which have "leap-frogged" to Settled Agriculture Phase 3. In some third­world countries, large ruminants feed on stubble in the fields following crop harvest ,nd consume straw. Some of their manure fertilizes the field but the farming system is predominantly crop-based with little integration.

It used to be thought that pastoral nomadism preceded settled agriculture but it is now considered to have been derived from it (Giigg 1974). Nomads live for he most part in arid and semi-arid areas where only grassland is present because of water shortages, ind aquaculture has limited potential.

Two major trends between 1300 and 1800 led to the development of the "mixed farming"characteristic of much of western Europe and the eastern USA from about 1850-1945: 1. the reduction and final elimination of the fallow and 2. pasture cultivation, in rotation with crops,

7 which provided feed for livestock. Nitrogen-fixing legumes were sown in mixed pasture with grass and helped to restore soil fertility as well as resting the soil for future cereal poduction.There was a gradual increase in the importance of livestock after the end of the Middle Ages,which was accelerated by the rise in real incomes in the latter part of the 19th century due to theIndustrial Revolution, and an increased demand for livestock products (Grigg 1974). Thisfarming system is called Settled Agriculture Phase 2 (Integrated Crop-Livestock). Theintegration of crops and livestock is a major feature of mixed fanning. Grass is cultivated as pasture, either permanently or in rotation with crops. A variety of crops is grown, particularlycereals, a large proportion of which is fed to animals on the farm or sold to feed mills. In Europe,root crops are also grown and some are fed to pigs and cattle. Livestock products (milk, butter,cheese, beef, poultry, pigs and eggs) are usually a more importart source of income for farmersthan crop produce. Livestock production is clearly integrated wikh arable fanning becauselivestock feed on crops grown on the farm, graze the pasture, and their manure ielps to maintain soil fertility (Grigg 1974).

Although agriculture has a history of at least 10,000 years, technical change was remarkablyslow until the middle of the 19th century. Western European agriculture became progressivelymore intensive from 1850 by using better seed, more fertilizer and mechaniz.tion. Although thetrend towards Settled Agriculture Phase 3 (Industrial Monoculture) started about 1850, it is onlysince World War II that traditional mixed fanning with integration of crops and livestock anddiversity of products has been replaced by specialization (Grigg 1974). The agriculturalrevolution in the West is becoming largely dependent on industrial inputs derived from scienceand engineering which is making farming more independent of the natural environment(Duckham 1959', 1966). The various components of industrial monoculture of modern agricultural technology are: improved varieties, chemical fertilizers, pesticides, herbicides,mechanization, feed concentrates, pelleted feed and pharmaceutical chemicals. Such resources are scarce a:d expensive in most third-world countries. In the face of increasing technicalcompl xity, a plethora of scient.fic and industrial inputs, rising labor costs, and the costadvantages of large and specialized farms, there has been a tendency in the West to "simplify"the farming system by reducing the number of enterprises (Duckham 1959, 1966). Farms now are less mixed and many raise only a single product. Modern farm economics have made theneed for integrated farming largely redundant in much of the West but there is growingrealization that industrial monoculture has much greater adverse environmental effects.

Raising livestock in confinement on feedlots, for which all or most cf the feed is purchasedoff-farm with a total separation of livestock and crop production systems, is a good example ofmodern fanning technology. Raising livestock in feedlots is fundamentally different from the"cut-and-carry" practice in which livestock such as small ruminants or rabbits are "stall-fed" withfeed obtained on or near the farm. In the latter, livestock feed comprises a farm subsystem.Feecdlot livestock production has been introduced to third-world countries (often on a "turn-key"basis) by vertically-integrated, agro-industrial companies but its benefits for small-scale farmers are questionable.

The organized cultivation of tree crops in plantations was introduced into third-worldcountries by Europeans and may also be considered as modern farming technology be.ause itgenerally involves monoculture to provide raw materials for agro-industry.

Most aiquaculture in the West and Japan is also industrial monoculture because, in general,single species are raised on pelleted feed and the systems are supported by mechanization (Edwards 1980)

Classificationof FarmingSy 'tems

A schema is presented to explain the concept of integrating fish culture into existing farmingsystems based on climate, type of water, water supply and size of farm holding (Fig. 3).

The greatest potential for integration of tropical aquaculture and agriculture lies infreshwater because most agriculture depends on fresh- rather than saltwater. However, some

--

All Faiming Systems

7 Climate Tropics Nontropical

Type of Water Freshwater Brackishwater

I __ _ Water Supply IrriqateJ Pinfed

F I Farm Size Small-Scale Large-Scale

Farming System Crops Crops/Livestock Livestock

-­1 1-4 Species Rice Maize

I Ruminants

1 Monogastrics

4 Large Small Pig Poultry

(Cattle, Buffalo) (Goat, Sheep)

Fig. 3. A classification of fanning systcrns.

brackishwater systems may be considered because rice is also grown in water which can be saline, either seasonally or throUtghout the year.

There is a need to conh:entrate future programs on rainfedrather than irrigated areas because the lbrmer coTnstitute 70-81)% of agricultural land and have generally been neglected in development projects. AquaCtilture also has great potential in irrigated areas where water stored and channelled to farmers is no longer a constraint in agriculture. However, much of the fish production data generated from successful integrated farming systems in rainfed areas would also be applicable to irrigated areas in a similar climatic zone. There is considerable potential for water storage in rainfed areas thro,:gh the efforts of the farmers themselves, in contrast to governmental provision of irrigationi. Water can be collected and stored for agriculture in areas with variable topography and seasonally heavy rainfall, with aquaculture in the storagereservoirs or tanks. Reservoir construction costs could be amortized in part by the simultaneous use of the water for fish culture. A fishpond is itself a method of water storage and may be used as an emergency water supply for other subsystems on the farm; for example, drinking water for livestock and a water supply for rice nursery beds.

Th : 'uture research and development focus should be on small-scale systems because large­scale f"'ms can usually alttract sufficient funds to develop aquaculture without external assistance. The concept of "sniall-s, 'e" is relative and depends to a large extent on the degree ofaridity. The minimum, faml size required to support a family is inversely proportional to amount and seasonal distribution of rainfall. A small-scale farm could be less than 1 ha in a TropicalRainy Climate whereas a small-scale livestock farm could be 1,000 ha in a Tropical Semi-arid Climatt' with insufficient rainfall to support productive pasture. The. term "small-scale" is used in

9 this study to emphasize that the target Is vil,ave level or lower (including family and groupoperations) and not large-scale commercial operations. A fish pond on an estate farm to augmentthe diet uf far-m workers may be regarded as a small-scale farm within the larger estate farm. Anincrease ir integrated i'arming might not increase the chances of employment on farms for thelandless because of widespread un- and under eiflp)oyment on small-scalc farms but it couldbrimng nutritional benefits and shuhld lead to the creation of more jobs in food processing and markeing.

Attempts have been made to develop agroclimatic maps based on agronomically significantparameters to compare areas of similar agrochimatic conditions and to establish data on theprod uctivity to be expected (IRRI 1974; Oldeman and Fr~re 1982). Temperature is not usually alimiting factor in the tropics and the duration of the growing season of crops depends on rainfall,except in areas with contrclled irrigation. Farmers traditionally adapt their cropping pattern tothe prevailing distribution of precipitation o,,:.- the year. A workshop at the Internationai RiceResearch Institute (IRRI) established eight agroclimatJc zones for rice-growing regions inSoutheast Asia (IRRI 1974). Agroclimalic zones were based on tic monthly rainfall and thelUmber of wet ronths with 200 mm or nilor% rainfall. The possibilities of growing two rice cropsare limited if there are less than five consecutive wet months. If there are more than nineconsecutive wet months, the Southeast As'i:mn farmer is mo't likely to grow two crops of puddledrice. Niaps from an FAO Agroecological Zones Project incorporate not only climatic variablesbut also c;nstraints imposed by soil,: (Old-:1n1w and Fr&re 1982). A constraint to both theseclassification systems is that they concentrate entirely on rainfed ag'riculture and do not consideradditional sources of water such as water from rivers and run-off. The maps are also crop­specific and do not indicate potential cropping pattern options. However, the approach is a usefulone and efforts should be made to delineate agroclin-atic zones for various integrated faningsystems to assist future agricultural development in the tropics.

Characteristicsof Asian Farming,Systems

Much more is Known about Asian han African farming systems from the point of view ofintegration of aquacultlure With agriculture The (- -hasis in this study is therefore on Asiansystems with the expectation that better understanding of crop-livestock-fish integration in Asiawill point the way to similar developmnentts in Africa.

Most small-scaie farmers in the third world may be characterized as Settled AgriculturePhase I wi:lh crops dominant and not integrated with livestock as in the traditional Europeanmi ixed f:mi eg system (Grigg 1974). An importawit premise of this study is that major increasesin farm productivity and profitability can probably be made by moving such farners lip toSettled Agriculture Phase 2 through development of integrated crop-livestock-fish farningsystems. The characteristics of a typical Asian farming system are outlined below as an example.A typical Asian farmer lives with his wife and four children on a fam of approximately 1.5ha and raises mainly rice. The family -wns one or two draught animals and raises severalchickens and ducks (IJoque 1984): land tenure and ownership/tenancy arrangements are vtuiable.There may be several enterprises of varying importance (Terra 1958; Webster and Wilson 1966;Grigg 1974) (Fig. 4):

1. wet-rice, almost always present on farms in tropical Asia and often the principal feature of the farm;

2. multiple cropping of other annual crops with rice in the paddy fields;3. permanent cultivation of annual or perennial crops, including staple roots and tubers on

dry land (upland);4. a mixed garden around the farmstead where fruits, v"getables and root crops are gr'own;

and 5. livestock, cattle or buffaloes kept mainly for draught, and scavenging poultry an./or pigs.

10

(d e s ad u lt i pl e ( Plan tation

Mixed Fields

Guenh t Crs Feedlot Livestock'

Livetoc irCd ro ti Wet

In ~~~~ ~~AsiaainErp(ireon tinrain Paddty dniy ariual na ra

ares)thee s atedeny orMixed gar ndoocp nicesn rprino h oa

oMarket Asian rayGarden Fig. 4. A traditional fann comprise several subsystems. Market gardens ev tlvej troi r owditiQnalmixed ga rdens but plan tation and fedlot l vestock fa son issyste were introduced frot the W a is.

In Asia, as in Europe (in response to increasing population density, particularly near urban areas) there is a tendency for mixed gardening to occupy an increasing proportion of the total arable land on the farm and evolve into market gardening, the growing of vegetables for sale (Terra 1954).

Farming throughout the humid tropics is characteri byaldthe neglect of livestock productivity. Most of the livestock in Southeast Asia are raised on small-scale farms rather than in commercial operations. Buffaloes and cattle are raised primarily as draught animals althoughthey are slaughtered when their working days are over. During the rice growing season the buffaloes and cattle subsist on rough grazing of poor quality and in the off-season for ipcethey gaze on poor quality feed in Daddy fields (Javier 1978). Cattle raising for dairy purposes is traditional in the Indian subcontinent but not in Southeast Asia, although there is now increasinginterest in dairying among small-scale farmers in the latter region (Grigg 1974; Javier 1978).

Fish culture, contrary to popular belief, is not at all widespread in tropical Asia. Farmers have captured wild fish from rice fields since time immemorial but it has been estimated that less than 1%of the irrigated ricefields in Southeast Asia are used to culture fish (Coche 1967) and little has changed since this estimate. There are a few traditional aquaculture iystems in tropicalAsia; for example, the Indian majnsocarp polyculture system which until recently was not an example of integrated farming because it was extensive with no fertilizer or feed inputs (Tripathiand Ranadhir 1982). Overseas Chinese were largely responsible for importing integrated fish farming technology to Malaysia, Singapore and T1. land, and perhaps Indonesia (Terra 1958).There has beer, a significant increase in the development of integrated crop-l ivestock-fi sh farming systems in tropical Asia over the past two to three decades but it still probably involves less than 1%of the small-scale farmers in~ the region.

Overseas Chinese farmers in Malaysia and elsewherc are involved in more intensive cultivation -of crops and livestock and sometimes fish. The livestock kept by Chinese farmers,mainly pigs, are well-integrated into the farming system. They are fed mainly on cassava and crop residues such as waste vegetables and rice byproducts and their manure is applied either to crops or to fishponds (W' bster and Wilson 1966). However, most small-scale farmers in Asia may be characterized as Settled Agriculture Phase 1 with crops dominant and not integrated with livestock or fish.

Research Framework

General Considerations

This study recognizes the value of a farming systems research approach for third-world agriculture development (IRRI 1982; FAO 1984; Gartner 1984).

There is an urgent need to identify and evaluate through on-station research and adaptivefield trials a range of technological packages of integrated farming systems involvingaquaculture that are suitable for the small-scale farmer, These should clearly be linked toopportunities and problems iafarming communities. An attempt is made in the following sectionof this study to identify strategies towards this objective by consideration of crop, livestock andfish farming -Libsystems for a tropical integrated farm. It must be emphasized at the outset thatthe tropical aquaculture database is so weak compared to those of tropical agronomy and animnalhusbandry that s,,icatific experimentation is still required to improve technology for seeminglybasic aspects of fish husbandry. Furthcrmore, it is of paramount importance that researchersstudy existing farming systems in detail for potential sources of pcnd inputs. Such an approachmust include in-depth socioeconomics research, the methodology of which is beyond the scopeof this study.

Crop Subsystems

For farms which currently have only crops (which may be grown in multiple croppingsystems) the basic question for an integrated farming syslem incorporating aquaculture is therelative values of crop byproducts such as green manure/compost for the cropping subsystem or as green manure/compost for pond fertilization and/or supplementary feed for fish.

Emphasis should be placed on the major energy crops that dictate the byproducts available on the farm. The major energy foods in the tropics are rice in wet areas, maize in dry or uplandareas, and to some extent sorghum where there is less -:tin. Rice is generally more important thanmaize in Asia and the reverse applies to Africa.

In the parts of tropical Asia where rainfall exceeds 1,000 to 1,200 rm/year, croppingsystems are usually based on rice. Rice is grown at the peak of the rains because rice is the onlycrop that tolerates flooding. However, it may be possible to plant upland crops such as maize,mungbean, cowpeas, and sweet potato in mixed cropping systems at the end of the rains to utilize residual moisture (Beets 1982).

Multiple cropping has great potential for increasing agricultural productivity. It is defined asgrowing more than one crop on a piece of land in a year and can take several forms: mixed orintercropping (planting more than one crop on a piece of land at the same time), relay cropping(planting crops in an already established crop) and sequential cropping - double or triple(growing more than one crop on a piece of land at different times in the yeai). The potentialproductivity of a multiple cropping system is given by the multiple cropping index (MCI):

MCI = crop area for1 year x 100%

cultivated area for 1 year

11

12

In multiple cropping systems, there is usually a certain optimum proportion of the species inthe mixture for dietary, economic, or agronomic considerations (Beets 1982). There is also thequestion of whether the crop is grown primarily for human, livestock or fish feed.

Fodder crops, particularly legumes could be introduced into existing cropping systems as intercrops, relay crops, or sequential crops without upsetting the regular cropping system (Javier1978). Various strategies have been discussed by Javier (1978). Aquatic grasses could be grownin ricefields as a fodder crop as an alternative to rice. There are several other possibilities with upland crops. More maize than normal could be sown and the excess thinned-out for fodder. Fastgrowing legumes such as soybean, mungbean and pigeon pea could be intercropped with maize apd a crop harvested before the maize closed-out the light. There are native species of SoutheastAsian legumes with known potential for livestock feed, e.g., Pueraria and Desmodium but also exotics such as various species of Stylosanthes (Schultze-Kraft 1986). It is possible to sowannual Stylosanthes on ricefield dikes to provide seed which can be broadcast into the stubbleafter harvest and grazed with volunteer weeds (Perkins et al. 1986). There is a vast array of treesand shrubs that serve as animal fodder, many of which fix nitrogen; for example Leucaena leucocephala (Brewbaker 1986). Tropical forage grasses have been studied for only a few decades but there are many species available (Mclvor and Chen 1986).

Livestock Subsysiems

The greatest potential for integrated farming systems with fish probably lies with mixedfarms having crop and livestock subsystems because livestock manure is a useful pond input.Farms that have only livestock fall into tvo distinct categories: 1. ranching farming systems insemi-arid and arid areas, with little potential for aquaculture because of water constraints, and 2. feedl,,,s.

Intensive feedlot livestock farming may be conveniently integrated with fish but it is acapital-intensive operation because of the purchase of feed off-farm, usually from agro-industry,and has limited relevance for small-scale farmers.

Recognizing the major use of livestock as draught animals, Javier (1978) posed the question: "with the increasing emphasis on mechanization in agriculture, will this mean thatthere will be no place for livestock any longer?" However, the place of livestock on small-scalefamas is assured, even with the advent of mechanization. It is feasible to grow three to five cropsort the same piece of land each year with increasing cropping intensities. Sixty percent of theannual production of digestible feedstuff is not useful directly as human food and livestock are apotentially attractive means of utilizing various kinds of crop residues (Javier 1978). There is avoluminous literature on the subject of nonconventional feed resources for livestock,summarized by Devendra (1985), which can also be referred to in the development of supplemeihtary feeds for fish.

Several reasons have been given for the relative neglect of livestock on 'small-scale Asianfarms; for example, 1. few farmers can spare land to grow fodder because it is needed to growcrops for human food; 2. the ti pical climate supports grasses of only poor nutritional value; 3.the tropical climate reduces the growth and fertility of cattle; a,-d 4. the tropical climate causes disease (Grigg 1974). However, these constraints are more apparent than real and can beaddressed with the knowledge available today from countries like Australia, which has highlydeveloped livestock farming systems in the subtropical and tropical areas of its landmass.

Most potential feed resources on the farm are in the form of crop residues such as rice strawand m,.ize stover which have low digestibility (Javier 1978). These are probably best fed toruminants (rather than used directly as fishpond inputs) and the ruminant manure used as a pondinput. Livestock can gain weight when fed on low quality crop residues supplemented with concentrates. There is often a seasonal distribution problem with livestock fodder. Rice straw,maize stover and empty legume pods are dry and .an be stored relatively easily but green fodder may need to be conserved by drying, ensilage, or composting (Javier 1978).

13

The highest fish yields from integrated farming systems have been reported from pondsreceiving feedlot livestock manure. The livestock had received high quality feed and their manure therefore had a high nutrient content (see below). Since livestock are a key component in a productive integrated farming system with fish, a major research effort is required to developtechnology for increasing the quality and quantity of livestock feed produced on small-scalefarms so as to increase livestock production directly and fish production indirectly in integratedfarming systems.

Ruminants, particularly large ruminants such as cattle and buffalo, have particular relevancefor the development of small-scale integrated farms because they are in widespread use in Asia as draught animals. However, many agricultural societies in Africa have yet to introduce theplough and therefore do not keep livestock for draught purposes. Ruminants can process fodderwhich is indigestible to humans but research is needed to identify strategies to upgrade thequality of their manure as a pond input because ruminants grazed on rough pasture and/or stoverhave manure with a low nutrient content. Small ruminants (sheep and goats) are normallyconsidered to be aninals of arid areas but they are important in certain areas in the humidtropics. Small ruminants are sometimes stall-fed using the "cut and carry" system. Free rangingrurninants, both large and small, are often paddocked at night which also facilitates manurecollection as a pond input. The collection of nitrogen-rich urine (as well as manure) as a pondinput has imnportant potential. For stall-fed ruminants, this resource is usually wasted, but trials are now beginning on the use of cattle urine as a pond input in India.

Monogastric livestock (pigs and poultry) with dietary requirements more similar to humansthan rumi'iants are often raised .i small numbers on small-scale farms as scavengers. The manure of feedlot pigs and poultry raised on commercial formulated feeds is high in nutrientsand is 1.valuable fishpond input. However, research is needed on the on-farm production ofnonconventional feeds for pigs and poultry and on tb nutrient content of their manure when receiving such feeds.

Fish Subsystems

Straiegies for increasing production

Two broad strategies for increasing agricultural production are to increase the fanned areaand/or to increase th(. yield per unit area (Grigg 1980). It is generally considered that there islittle potential for increasing the area of arable laid, with the exception of Africa. The expansionof the area devoted to aquaculture must therefore be considered as a major strategy because thereis relatively little aquacti'ure in the tropics at present, including Asia, and high yields of fish canbe obtained from small areas comipared to those required for significant production of mostarable crops. Fish culture is a highly attractive option for increasinz, the production of highquality animal protein.

Increased yields should also be targetted, preferably by the more widespread application oftraditional agricultural technelogy involving pond inputs generated on-farm (integrated farming)as opposed to the adoption of modern agricultural technology whh its dependence on agro­industrial inputs.

Extensive, semi-intensive and intensive aquaculture

The degree of intensification of fish farming is defined according to feeding practicesbecause :hL tsually comprise more than 507c of the total operating costs in an intensive system. However, intensification is associated with increasing usage of capital, labor andmechanization. A useful classification is as fllo,

14

1. Extensive systems utilize natural feed produced without intentional pond inputs. They are excluded by definition from an integrated crop-livestock-fish farming system with the exceptionof certain integrated rice-fish fwTning systems in which fish may derive benefits from inputs added solely for rice cultivation.

2. Semi-intensive systems rely on fertilization to produce natural feed and/or supplementaryfeed (but with a significant amount of the fish diet supplied by natural feed) and are typical components of integrated crot-livestock-fish farming systems.

3. Intensive systems have all the fish nutritional requirements provided by a nutritionallycomplete pelleted feed with little or no nutritional benefits from natural feed produced in the pond. Trash fish, a byproduct of capture fisheries, is also used as feed in certain intensive aquaculture s",ems. Such intensive aquacilture would normally not occur in a crop-livestock­fish farming sy.tem because it is difficult to formulate and produce a nutritionally completepelleted diet from ingredients produced only on the farm. Most western and Japaneseaquaculture systems fall within this category.

Essentiallv, fish raised with I vestock in an integrated fanning system feed in the seni­intensive mode. A major part of their nutrition is derived from natural food which develops in the pord duc to fcrtilization of the water by manure and fish feces. However, the relative contribution of natural feed to fish nutrition decreases as the quality and quantity of supplementary feed increases.

Fish yields from aquaculture systems range over three orders of magnitude (Fig. 5):

Yield (t/ha/year)

:[~~~ o,0_ 201_o] 05100-1i oo_.oo I I

I'l l II

I I

=Intensive ]" II

I I

No Nutritional Inputs I

Low QualityManure,

High QualityManure

I High QualityManure, I

Pellets, Aeration,

Pellets, Raceway

Macrophytes Pellets, Recirculation I I I

I II

Aeration I

Fig. 5. Intcnsifica.iti of 'AJacuutire systerns,

1.0-1 t/ha/year from extensive systems with no nutritional inputs;2. 1-5 t/ha.'year front semi-intensive systems with low quality manure and/or macrophytes as

supplementary feed; 3.5-15 t/ha/ycar from semi-intensive systems with a high quality manure input;4. 15-20 t/ia/year from semi-intensive systems with a high quality manure input and

pelleted feed inputs and aeration characteristic of intensive systems;5. 20-100 t/ha/year from intensive systems with pelleted feeds, aeration and water

recirculation; and 6. 100- 1,000 t/ha/year from intensive systems with pelleted feeds and running water (ponds

or raceways).The upper fisl, yield obtained to date in a livestock-fijh farming system with high quality

manure (duck or pig) as the sole pond input is about 10-12 t/haiyear from both small-scale experimental and larger-scale commercial systems (Olih 1986; Wohlfarth and Hulata 1987).This is impressive, but the system is essentially a "black box" because pond dynamics - the biological and chemical bases of production - are poorly understood.

15

It is proposed that an integrated livestock-fish farming system with high quality manure asthe only input be used for further basic scientific research to understand how the systemfunctions. The knowledge from such studies could then be used in attempts to increase the fishyield from integrated crop-livestock-fish farming systems in which the pond receive lower quality inputs such as low quality manure and/or macrophytes (vegetation) as supplementary feed.

Interactions in crop-livestock-fish integrated farming systems

Possible on-farm interactions between the various subsystems in a crop-livestock-fishintegrated farming system are presented in Fig. 6. The schema excludes products from the various subsystems and merely indicates on-farm link ages. Rice-fish culture is well-established in certain Asian countries (de la Cruz and Carangal, in press) and involves a vadiety of systems e.g., trenches and ponds, constructed in rice land (Plates I and 2). Livestock excreta (manure)

1.,t!. (,S hcnl I J hIt liiln intorictions bctm.ccin the '. ariou suhsysicuis ina c:rp-livcsit~ k-fai

... . ., -

Plate I. A peripheral trench surrounding a ricefield in a rice-fish Platc 2. A fishpond conltructed within i icefield in a,rainfed area in culture systern in the Philippines. The aquatic vegetasble taro Northeast Thrailand. Pond water is also used to irrigate vegetables (Colocosia esculen ta)is also cultivated on the edge of the trench. cultivated on the dike.

16

may be used as a fishpond input or to fertilize crops. It is also feasible to incorporate manure intolivestock rations. Human excreta may also be used to fertilize the pond or as a crop fertilizer.Crops may be fed to livestock or used as supplementary fish feed. Water from the fishpond maybe used to water crops (Plate 3) or as drinking water for livestock (cover plate). Mud removedfrom the pond may be used to fertilize crops. Fish that are too small to be marketed may be used as a high protein ingredient in livestock or fish feed. The concepts are essentially Chinese inorigin and several of the links are supported by a wealth of empirical farmer experience althoughthey have yet to be subjected to the rigor of s;cientific analysis.

~- A,

4/ ..

. . .. w_ "7t ui "'N?

plate 3. Dry seasonal cultivation of vegetables on a fishpond dike in Cover plate. Small-scale integrated crop-livestock-fish farming in aNortheast iThailand. rainfed area of Northeast Thailand. This rice farm has a small fishpond that provides fish, permits dry season cultivation of vegetables on the dikes and supplies drinking water for livestock.

Deriving fish culture practices from natural aquatic ecosystems

There are three fish culture systems that have been derived from natural aquatic ecosystems:1. Vegetatior.-fed systems in which terrestrial plants and/or aquatic macrophytes are fed to

fish (Plates 4a, 4b, 4c); large amounts of excreta are produced by the inefficient digestive processes of macrophyte herbivorous fish (Edwards 1987). These excreta act as pond fertilizers that produce natural food for plankton/detritus filtering fish (Plate 4d) and carnivorous/detritusbenthic feeding fish. Aquatic macrophytes growing in the pond are not part of the systembecause, in a well-nianaged system that has adequate inputs to support good fish growth, they are shaded out by phytop!ankton.

Plate 4a. Terrestrial vegetation (pumpkin leaves) being chopped-up Plate 4b. An aquatic macrophyte, duckweed, being used as fodder forprior to use as fodder for fish in Northeast Thailand. fish in Central Thailand.

17

Plate 4c. (;r:is carp raised in macroph)te fed pond in Ce ial Plate 4,1. Nile tilapia raised in a macrophyte-fed pond in Central'lloand. 'llailand.

2. Excreta (manurc)-fcd systems for plankton/detritus filtering fish and carnivorous/detritivorous benthic feeding fish (Plates 5a, 5b).

3. Trash fish-fed syst ems for purely carnivorous fish such as the culture of snakhead(Channastriata)and walking catfish (Clarias spp.) on byproducts from the trawling industry in Thailand.

Plate 5a. Small-scale integrated chicken-fish system in Plate 51, Small-scale integrated pig-fish systen in\Vet Java, indonesia. Northeast "lhailand.

A simplified diagram of vegetation-fed and excreta (manure)-fed fishpond systems ispresented in Fig. 7. System 3 is omitte as it is not , viable option for most integrated farms dueto the high cost and limited avail ' "..y of the feed input. Such farms require more energeticallyefficient fish that feed lower,..,, the food chain.

Choice of fish species

A total of 31 species is listed in Table I as potential candidates for integrated fanningsystems involving aqLuaculture in the tropics. All are native to Asia with the exception of thetilapias. The list of predacious species could be expanded to include some African species butpredacious fish are really of less interest in integrated farming than those withplanktivorous/herbivorous/detritivorous trophic niches. Predacious fish have been used to controlthe recruitment of tilapia but this role will likely diminish with the introduction of monosextilapia culture. Sixty-five per cent of the fish listed are native to the tropics; the remaining 35% are native to wam-temperate/subtropical zones but thrive in the tropics.

18

Ex reta Vegetation1311r ..... ..... .... .... ... .I .. .... ..... .... .... ... .. .. .

7 Herbivorous Fish

IExcreta Physoplankton Zooplankt-~,," ' l _i. Fi.L a 7 o c n - - Fret

Excreta MDetritus i Nn et s Plankton/Detri tus Filtering Fish

-A-­Carnivorous/D)etrital Excret

p d tiv a n r as ie p l l D etritivores m r

Fig. 7. Food chains in vegetaticxl and excreta (manure) fed fishponds. Solid lines represent pathways of particulate matter and broken lines soluble nutrients fromn excreta. Major components and pathways are indicated by thicker lines.

Polyculture, the culture of more th~an one species of fish together in the same pond, hasgenerally been regarded as more productive than raising individual species separately (monoculture). The rationale behind polyculture is that fish have different trophic and spatialniches and that with polyculture a balanced fish population, with different species thatcomplement each other, can occupy all the niches in the pond. However, there is surprisinglylittle experimental evidence to Support the concept of polyculture with the exception of theincorporation of planktivorous fish in the traditional European monoculture of the common carp(Cyprinus carpio),which is a benthic feeder (Opuszynski, 1981; Yashouv 1971). Trophic nichesof fish overlap to a much greater extent than is general!y appreciated and Nile tilapia(Oreochromisniloticus), which occupies several niches, has yet to be evaluated againstpolycultures under the same set of experimental conditions. 0. niloticus is an exceptionallyversatile feeder. A recent study in the Sudan (Hickley and Bailey 1987) described it as takingphytoplankton in the water column, periphyton and fine particulate organic matter from plantand other surfaces, and benthic organic detritus.

The fish in Table I are listed according to major (**) and minor (*) trophic and spatialniches (+). The niches of the various species need to be established under variousfertilizer/feeding regimes. The competition index (C) of Reich (1975) couid be used to quantifythe performance between species in polyculture:

C =(A - B)/A x 100%

where A = yield of a certain species in monoculture;

and B = yield of the same species in polyculture with a second species.

If B < A and C is positive, there is competition between the two species.If B > A and C is negative, the presence of the second species increases the yield of the first.

Table 1. Fish spec cs as potential candidates for tropical crop/livestock/fish farming systems according to major () trophic niches. Minor trophic () and spatial +) ni,hes are also indicated.

Scientific nani

He/ostoma temmincki Trichogaster pectorali, Chanos chanos Oreochromisaureus 0. mossambicus

0. ni/aoticus

Hypophthamichhys molitrix

Labeo rohira

Osteochilus hasse!tii

Aristichrhys nobilis

Carassiuscarajsjius Catlacatla

Osphronemus goramy Tilapia rendalli

T. zillii Ctenopharyngodon idella

Megalobrama amblycephala Parabramis pekinensis Puntius gonionotus

Cirrhinus molitorel/a

Cirrhinusmriga!a Cyprins carpio Mylopharyngodon piceus Mugil cephalu; Macrobrachium rosenbergii Pangasius pangasius

Channa striata

Clariasbatrachus

C. m.ncrocephalus

Lateolabrax japonicus Lares ca/carifar

lainily

Anabantidae

Clupeidae Cichlidae

Cyprinidae

Cyprinicae

"'aria

Anabantidae Cichlidae

Cyprinidae

Cyprinidae

Mugilidae Palaemoni-ae Schilbe;dae

Channidae

Clariidae

Serranidae

Common name Filter feede, Trophic niche i Macro-particilate

Spatial niche

Phytopiankton Zoopiankton Macrophyte Detritus/ Predacious Surface Bottom Column Invertebrates

Kissing gourami Snakeskin gou rami

*neba

I Milkfish Blue tilapia Mozambiqu? tilapia

* * *.+ +

Nile tilapia

Silver carp *+ Rohu Nilem

**+ +

6ig head carp . t + Crucian :arp

+

ui;,nt gourami

Grass carp Wuchang fish

*+

Chinese bream Silver barb

**

Mud carp I+ Mrioal Common carp Bmack carp +

Grey .nuilet + Giant Ireshwater pravn Silver striped catfish ___"_+

I +

Snakehead J {-T Walking catfish

Sea perch Sea bass I _.

20

There is a near infinite number of potential polyculture systems considering the largenumber of potential species in Table 1.There are 11 permutations of two species in monocultureand polyculture, with species ratios of 1, 2, 3 and 4 (Table 2). A two species polyculture system

Table 2. Number of permutations of two species in munocuiture, and in polyculture with four ratios.

Species ratios Number of species Species Species

(species ratio) 1 2

(a) Monoculture

(2 permutations)

11:0) 10

(b) Polyculture(11 permutations)

2(1:1) 1 1 2(2:1) 2 1

1 2 2(3:1) 3 1

1 3 2(3:2) 3 2

2 3 2(4:1) 4 1

1 4 2(4:3) 4 3

3 4

involving a plankton/detritus column feeder and a carnivorous/detritus benthic feeder might beappropriate for an excreta (manure)-fed system. There are 91 permutations of three species inmonoculture and polyculture with two species, and polyculture with three species, with speciesratios of 1, 2, 3 and 4 (Table 3). A three species polycuhure system involving a macro'phyteherbivorous fish, a plankton/detritus column feeder, and a carnivorous/detritus benthic feedermight be appropriate for either a macrophyte-fed system or a macrophyte and excreta (manure)­fed system. Two species would be required for systems that receive only manure.

The number of permutations soon becomes impossibly large with more than three species inpolyculture. To evaluate even a limited number of permutations would clearly need a majorresearch effort with large numbers of experimental ponds, personnel and equipment items.Polycultures of at least six species are common in China and India (Lin 1982; Tripathi andRanadhir 1982) but there is a dearth of scientific data to support the use of such large numbers of species.

it is recommended that research be conducted initially with polycultures of two species formanured systems and three species for systems in which macrophytes are used as supplementaryfeed, taking into account the market demand for these species. The versatile plankton/detritusfeeding Nile tilapia (Oreochromisniloticus) and the versatile bottom feeding common carp(Cyprinus carpio)are suggested for both manured and macrophyte-fed systems with the additionof a suitable macrophyte feeding fish for the latter system. The best options for the macrophytefeeding fish are grass carp (Cenopharyngodonidella), silver barb (Puntiusgonionotus)andTilapiarendalli, according to local circumstances.

21 Table 3. Number of permutations of three species, (a) in monoculture, (b) polyculture with two species and (c) polvculture with three species, with four ratios.

(a) Monoculture (3 permutations) c) Polyculture with 3 species (55 permutations)

Numbe, of Number of Number ofspecieF Species ratio spccies Species ratio(species Species Species Species specie Species ratio(species Species Species Species (species Species Species Speciesratio) 1 ratio) 1 2 3 ratio) 1 2 3

1(1:0) 1 - - 3(1:1:1) 1 1 3(4:3:2) 4 3 2 -- 1 1 3(2:1:1) 2 21 1 4 31 2 I 3 4 2 1 1 2(b Polyculture with 2 species 3 2 43(3:1:1) 3 1 2 2 4 3(33 permutations) 1 3 1 2 3 4

1 1 32(1:1) 1 1 4:1:1) 41 1 4

1 ,

3(4:3:3) 4 3 31 1 3 4 31 1 1 1 4 4 42(2:l1 2 31 - 3(2:2:1 2 2 1 3(4:4:3) 4 4 32 1 2 1 2 3 4 41 1 2 2 41 2 - 3(3:3:1) 3 3 1 3(4:4:1) 4 4

3 4 11 2 3 1 3 1 4 41 2 1 3 3 4 1 42(3:1) 3 ­1 3(3:3:2) 3 3 2

3 - 1 3 2 31 3 2 3 3

3 3(3:2:1) 3 2 1 3 1 3 1 2 1 3 2 3 12(3:2) 3 2 ­ 2 1 3

3 2 1 3 2 2 3 ­ 1 2 3 2 - 3(3:2:2) 3 2 2 - 3 2 2 3 2 - 2 3 2 2 32(4:1) 4 ­1 3(4:2:1) 4 2 1 4 1 4 1 2 1 4 ­ 2 4 1 1 - 4 2 1 4

4 1 4 2 1 4 1 2 22(4:3) 4 3 - 3(4:3:1) 4 3 1

4 4 1 33 4 3 4 1 3 - 4 3 1 4 - 3 4 1 .1 3 - 4 3 1 3 4

Fishpond dynamics

The objective in a manured fishpond is to fertilize the water to produce enough natural foodfor the fish but not an overabundance of plankton, particularly phytoplankton, that can affect fishgrowth or survival by adverse environmental conditions. Both autotrophic and heterotrophicfood chains proceed simultaneously in a manured pond, involving phyto-, zoo-, and bacterio­plankton as well as benthic invertebrates and bacteria in the sediments, although the extent towhich the various natural food components are exploited by fish depends to some extent on thetrophic and spatial niches of the fish community (Colman and Edwards 1987).It is desirable to determine not only the biomass or standing crop of the various types ofnatural food but also their productivities so that their potential relative contributions to fishnutrition can be assessed. Ideally, it would be desirable to isolate the different components of thefood web to assess thc'r,feed value but this is difficult to do in practice. The maximum sustained

22

rate of phytosynthesis in fishponds in the tropics is about 4 gC/m2/day (8 g of biomass/m2/day) or equivalent to about 30 t dry weight of phytoplankton/halyear (Colman and Edwards 1987).Assuming a feed conversion ratio (FCR) of 2:1 (dry phytoplankton to wet fish), the maximum fish yield would be about 15 t/ha/year (30 kg/ha/day), which is close to the maximum reportedyield from ponds loaded with high quality livestock manure. Research is needed to assess fish growth on different types of phytoplankton because there is convincing evidence that hlue-green algae are more digestible (especially by tilapia) than green algae (Colman and Edwards 1987).Methods to introduce and sustain blue-green algae (blooms) require further study, possibly inolving seeding. The competitive interactions amongst different types of algae in fishponds should also be studied. Zooplankton are widely acknowledged to be an important natural food for fish, particularly for fry. Bacteria whi::h can be entrapped by mucus secretions of both tilapiaand silver carp may be an important source of nutrition. Much more research is warranted on feeding pathways in manured ponds and on the mechanisms by which fish filter and digestplankton and p:arculate matter.

The nutrient dynamics of fishponds, particularly with respect to C, N and P require elucidation. To maintain a daily photosynthetic rate of 4 gC/m2/day in a 1-m deep fishpondwould require minimum daily inputs of 4, 0.8 and 0.08 g C, N and P/m 2/day, assuming that the C:N:P ratio of phytoplankton cells in a light-limited pond with excess nutrients for growth is 50:10:1 by weight (Goldman 1979), and assuming 100% transfer efficiency and no nwtrient recycling within the system. The latter two assuiptions of course are incorrect but tend to cancel each other out. A better knowledge of nutrient dynamics within the pond ecosystem would enable rational decisions to be made concerning the amount and frequency of nutient loadings and whether these should be constant, increased, or decreased with time during the fish growth cycle.

A major consideration in manured ponds is water quality, particularly dissolved oxygen(DO). There are large diurnal fluctuations in DO ii, -1steady-state manured pond, due largely to the presence of phytoplankton (Fig. 8). However, p:oblems with low DO at dawn occur only if the phytoplankton are not growing because, on a 24-hour basis, they generate more DO tl.an they use in respiration if net photosynthesis is positive (Colman and Edwards 1987). Experiments on the tolerance to low DO of various fish species should be conducted in diurnally fluctuating as

24­

20 ­

16

a)

° 12 /

(n0 ....­

0j

6AM noon 6PM midnight 6AM noon 6PM midnight 6AM noon 6PM midnight 6AM

Day I Day 2 Day 3 Fig. 8. Diutal changes in dissolved oxygen (rg/I) in infertile water (dotted line), fertile wa'er (dashed and dotted line) and hyperfertilp water (solid line) in the tropics.

23

opposed to constant DO regimes. Such experiments can be carried out in the laboratory in clearwater systems but a more valid assessment needs to be conducted in outdoor systems withchanges in DO effected by the phytoplankton. Chronic sublethal effects of low DO, which canlead to reduced fish growth over long periods of time, should be studied in addition to short-term lethal effects.

Toxic products of organic matter degradation, particularly ammonia, nitrite and hydrogensulphide are probabl; n.t a problem in a steady-state manured pond system althoughcatastrophic inputs of organic matter from shock loadings of manure or from the collapse of algal blooms can lead to fish kills.

Types of nutritional inputs to ponds

The major measure of quality for a pond input is its C:N ratio. This applies to feeds as well as to manures because there is a highly significant correlation between the nitrogen content offood and its absorption efficiency by fish (Pandian and Marian 1985). Bomb calorimetry couldbe used to get a "common currency" for pond inputs and outputs so that meaningful efficiencies:f fish productivity could be developed for a wide range of inputs. However, nutrient value mustbe considered as well as energy. Carbon isthe single most important nutrient in biologicalsystems in terms of the quantity incorporated into organisms but nitrogen is usually the firstlimiting nutrient because (f its volatility. C:N ratios are at least two times less in high qualitythan low (qualityinputs.

There is a well-known r.-ationship between fish yield and various types of pond nutritionalinpu s (Hepher 1978; Van der Lingen 1959) (Fig. 9). Fish yield increases with an increase in thestatus of pond nutrition but to benefit fully from increased food availability it is necessary toincrease the density of fish in the pond. Pond experiments should therefore be conducted at -range of stocking densities: experience to date suggests hat 0.1, 0.5, 1, 3 and 5 fish/m2 would be a suitable range.

Manure + Cereal + Complete Feed

2 Manure

NoInu

Fish Stocking Density

Fig. 9. The relationship between fish stocking density and fish yield as a function ofvarious types of pond nutritional inputs. Modified after Van der Lingen (1959).

24

Recent studies have demonstrated high yields from ponds loaded with manure from feedlot livestock. Hopkins and Cruz (1982) obtained extrapolated net fish yields of up to 10 t/ha/year from 400-m 2 and 1,000-m2 ponds in the Philippines, using only pig or poultry manure, without inorganicfertilizer or supplernentaryfishfeed.Similar yields were obtained at the Asian Institute of Technology (AIT), Bangkok and in villages (Plate 6) in Central and Northeastern Thailand using duck manure as the sole pond input (Edwards 1983). A mean annual net yield of 175 kg of fish was obtained from a 200-M2 pond fertilized with the manure of 27 ducks in villages. It was estimated that this could supply almost all the annual animal protein needs of a farnily of five people.

Ft,

11"T

Plate 6. An Alladaptive field trial with do k-fish integration ii avillage in Cent ra I'Ih ia ml

However, the effects of the va:ious types of organic manures on fish yields remain to be adequately assessed. In particular, why manfure from feedlot livestock gives a Much higher yield than manure. from grazing ru1in11111tS. Although dlawn DOs were close to zero in cluck-manured ponds in the AlT stuidy, DO concentrations during the afternoon were as high as double supersa tuLra tion due to intense phytoplankton photosynthesis. A hypothiesis wVas made that fish prodcIItivity is directly proportional to manure N-content and that similar fish yields lo those obtained in the duck-mianured pond could be obtained by adding buffalo manure to provide the same N-loading rate (Edwards 1983). However, subsequent experimentation gave much lower yields from buffalo- than from duck-manured ponds. (AlT 1986). A higher dry matter loading rate of buffalo manuire Was used to obtain N-loading rates comparable to those for duck manure. ThL. Caused an adverse- DO regime in the pond. It appeared that the main oxygen demand was from the buffalo manuire itself and not from the night-time respiratory demand of phiytoplanktor. This cor-tnists with the situation found in poaids fertilized with high quality duck Manure.

It is part;cuLarly important to investigate the technical feasibility and economic viability of using inorgaric fertilizer supplementation to improve grazing ruminant ma-nure from a "low quality" to a 'high quality" manure to increase fish yields. Preliminary experimentation at AIT with inorganic fertilizer supplementation of buffalo manure has recently led to a significant increas. in fish yield from buffalo-manured ponds with enrcouraging gross margins for the cost of commercial fertilizer compared to the farmgate price of fish (AiT, unpub. data).

25

Byproducts such as cereal brans (Plates 7a and 7b) and oil cakes are already known to begood quality supplementary fish feeds but research is required to improve the nutritional value oflower quality feedstuffs, such as crop residues and straw. Essentially, there are three approachesto recycling low quality (high C:N ratio) byproducts in a fishpond:

1. aerobic composting on land;2. aerobic utilization by broadcasting chopped material over the pond surface so that it can

be directly consumed by fish or enter aerobic aquatic decomposition pathways; and3. anaerobic composting in situ by heaping the matter in the pond.

4 Haw 7a. Miai/e bran (nadya) used as a supplementary fish

- . . • I,- e.d r-

N- - - .-

Plate 7b.Rice bran feeding in l- una, Philippines.

The second strategy would probably be the most efficient - aerobic composting ordecomposition ii the pond itself - becaIse the loss of nutrients would likely be less by aerobiccomposting in the pond than on land. However, all three strategies merit further study.

Microbial preconditioning of plant matter by aerobic composting on land leads to areduction in the C:N ratio because C is lost and N is conserved. However, the N becomesincreasingly refractory with time as composting proceeds. It becomes 'locked up' in compoundsthat do not break clown easily (Pullin 1987). Short-term experiments should be conducted with acompost maturation period of days rather than months to correspond to the thermal maximum ofthe compost pile. This is when the microbial biomass should also be at its peak.

26

The efficiency of conversior. of farm crop residues into farm produce should be compared between a crop-ruminant system (in which manure is used as a crop fertilizer) and a crop­ruminant-fish system in which manure is used as a pond input. An analogy has been drawn between the rumen and the fishpond: low quality inputs in both are processed into higher quality microbial natural feed for target organisms (Schroeder 1980) althougl' this is a doubtful analogy (see discussion in Mo-iarty and Pullin 1987). Moreover, it may be more biologically and economically efficient to process low quality vegetation and crop residues (with appropriate pretreatments to improve digestibility) by feeding them to ruminants and then to use their rnaiiure as itpond input rather than to utilize these materials directly (with or without pretreatment or nutrient supplementation) as pond inputs.

Vegeta:ion, both terrcstrial and aquatic, needs to be assessed as a direct feed for herbivorous fish and as a fertilizer to produce natural food after incomplete digestion by herbivores. The direct-feeding value of aquatic macrophytes in particular might be improved considerably if their moisture content were reduced.

Complete feeds are normally dried and pelleted and are produced by agro-industry. Resea.rch should be conducted into the on-farm manufacture of low cost wet or dry pelleted feed using 'ow cost materials and methods of feed storage. A major problem is the replacement of fish mteal as the major protein source for commercial diets. Plant protein sources such as soybean can only replace a limited percentage of fish meal. Small fish that are too sniall to maiket, harvested from ponds or ricefields, may be a suitable replacement for fish meal. Other sources of anmal protein such as snails or tubificid worms need to be assessed as nonconventional animal protein sources of pelleted feed. A well-manured pond usually has abundant high protein natural food so it ray be feasible to feed fish with cheaper energy rich pellets to complement the natural high protein diet- However. such diets may need to incorporate feeding stimulants to be acceptable to fish (Mackie 1982). Betaine, amino acids, "arnino-acid like" substances and inosine and its derivatives have been characterized as feeding stimulants for teleost fish (Carr 1982). Research is required to identify low cost feeding stimulants, particularly those that occur in .m-fium products. Studies are also required on low cost methods of on-farm storage of pelleted feeds.

Physical characteristics of the pond

Fish yield may be influenced by the size of the fishpond, irrespective of the rate of fertilizer/feed inputs per unit area. The water quality may be better in a larger than a smaller pond because of the greater wind effect on the surface. Wind induced water movement is a major factor in water circulation in static water ponds. However, large ponds have proportionally less edge/marginal zone (perhaps the most fertile aie-. ota pond) than small ponds. This could adversely affect fish yield. Research is reouired to determine the optimal size of fishponds for different systems, commensurate with good pond management.

Ponds are commonly only 0.8-1.5 m deep in the tropics although ponds in China are usually 2-3 m deep. The optimal pond depth for the tropics remains to be detennined. The effect of pond depth on fish production could be assessed by computing pond inputs on an areal as well as a volume basis for both shallow and deep ponds. A factorial experimental design involving stocking density would need to be incorporated into such studies. Fish productivity might not vary with depth in a manured pond because the productivity of the natural food organisms in the water is a function not only of the nutrients contained in the inputs but also of solar radiation at the surface, which is independent of depth. However, deep ponds may lead to greater fish yields than shallow ponds if significant amounts of supplementary feeds are given. Furthermore, deep ponds may be needed in rainfed areas with seasonal rainfall to store water to permit fish culture during the dry season.

27

Pond sediments

There is considerable controversy concerning the role of pond sediments in fish production.The water column may have the more important role in productivity in a manured pond becauseit is three-dimensional compared to the two-dimensional sediment'water interface. Furthermore,the water column contains all the phytoplankton. However, bacterial produ, ,ity should be greatest ,!ihe sediment/water interface due to bacterial decomposition of sedimented organicmatter (Fig. 10). Thus, the sediments could be an even more important site of nutrient;v'-,_-ncration than the water column. It is hard to pirtition autotrophic and heterotrophic foodchains in a po':d loaded with significant amounts of manure but experiments should beconducted in which pond sediments are physically separated from the water column. The effects on fish yield. of various densities of benthic feeding fish as "biop,.rturbators" should also beassessed. Mechanical disturbance of the sediments or 'stirring' aLo merits investigation. A totalcarbon fixzation rate of 8 gC/m2/day has been reported 'or fi .,!pords with sediments regul,'-lystirred so that sedimented detritus and associated bac',,ia were resuspended in the highly aerobic water column (Costa-Pierce and Craven 1987). This is iconsiderably higher carbon fixation rate than those previously obtained from static, unstirred ponds. The strategy of sediment stirringmay be the key to elevating the fish productivity in,manured ponds. However, it should lot be

Atmosphere/Water Interface (Surface)

Wtater Column (Volum)

.A:::::::: .":: : . . . . . . ...... !, : '

~Sediment/.

Interfdce (Surface)

Fig. 10. A three dimensional representation of the fishplvd water colunn.

forgotten that the most eificient food chains are aerobic not anaerobic and that photosynthesis byphytoplankton is the major source of dissolved o,:ygen. A balance would need to be reachedbetween increased bacterial productivity from the resuspension of bacteria-covered detritalparticles in th(; aerobic water column (which would also increase nutrient regeneration andstimulate phyloplankton growth) and reduced photosynthesis due to water turbidity.

28

However, a "sediment-degradation" phenomenon has been reported from Israel (Rain et al. 1982) in which fish growth is inhibited 50-70 days after stocking. This has been attributed to the accumulation of organic matter leading to anaerobic conditions and the production of toxic H2S. Sediments are usually air-dried when fish are harvested by pond draining before a new cycle is started. This oxidizes the sediments, at least partially. However, there is no information on the effect of varying the drying period on the mineralization of sediments of different depths, nor of the effects of drying on water column productivity wher. the pond i5 refilled. In China, excess fishpond sediments are removed and used as crop fertilizers (Plate 8). Studies are also required on the value of fishpond sediments as fertilizers for terrestrial crops.

Plate 8. l'ishpond szdunents collected during the fish culture cycle fertilize mulberry cultivated on a fishpond dike in Guangdong, South China.

Research is clearly warranted on all aspects of the management of fishpond sediments during the fish culture cycle.

Fish stock management

Most fish culture comprises a single stock-single harvest operation in which fingerlings are stocked at the start of the cu!ture period and are harvested by pond draining for market at the end of the growth cycle (Fig. 11). The increase in weight of the fish in the pond follows a sigmoid curve: slow during the first phase, because the individual weights of fingerlings are small, and more rapid as the fish grow larger. The third phase is one of slow growth because the carryingcapacity of the pond is being approached. The carrying capacity may be defined as the total weight of fish in the pond that can be supported by the available feed resources and water quality(Fig. 11).

However, significant incre,,es in fish yield may be obtained by utilizing more fully the spatial and nutritional resources of the pond throughout the culture cycle. A higher weight of fish should be stocked at the outset to eliminate the slow weight increase of phase 1. An intermediate harvest should be carried out at the upper inflection point of the curve at the end of phase 2 when the increase in fish weight slows because the carrying capacity of the pond is being approached.Although there are as yet few experimental data to support this hypothesis, the cumulative harvests from such stock management may be at least double that in a single stock/single harvest cycle. Research en fish stock management could significantly increase the yields from most fish culture operations, not just integrated farming systems. It could also have enormous economic advantages for farmers; for example, a more even supply of produce to market and improved cash flow.

29 (a)Single Stock - Single Harvest

Final Harvest

ISingle

Harvest

Ph Phase 2 Phase 3

Time

b) Intermediate Harvests

Harvest 1 larvest 2 Final Harvest

I I Multiple Harvest

I I

Initial Weight I

I j

Time

Fig. 1I. Fish stock management to increase fish yields.

Systems modelling

Collection and analysis of data from complex aquaculture systems is difficult. Howeve,attempts are now being made to improve the format of data collection from fishponds (e.g.,CRSP 1986). Several groups have begun to apply powerful statistical techniques, such asmultiple regression and critical path analysis, to data sets from waste-fed fishponds, from on­station experiments and farms (e.g., Milstein et al., in press; Pauly and Hopkins 1983; Prein1985). Data sets can be analyzed and compared by different working groups anywhere in theworld. The key is the collection and formatting of data in an appropriate form and theavailability of suitable hardware and software. These techniques hold great promise for thefuture. A review of the application of systems modelling in aquaculture has been prepared byCuenco (in press).Systems modelling is also an important field of future research for integrated agriculture­aquaculture farming sy stems. Ecosystem modelling techniques have been applied to fishpondsby Cuenco et al. (1985a, 1985b, 1")85c) and Svirezhev et al. (1984). Modelling techniques haveyet to be applied to integrated farms having an aquaculture subsystem but they are powerfultools for elucidating the major factors that control productivity and profitability (the 'bio­economics' of the farm) and hence the choice of management options.

30

There is a wealth of experience in such data analysis and modelling techniques available from agriculturists, especially farming systems specialists, of which aquatic systems researchers currently know virtually nothing.

Potential health hazards

The introduction of a fishpond as a farm subsystem should not pose any unacceptable risks to public health. There is a possibility that livestock manured ponds may present health problems for humans because some diseases of animals are transmissible to human beings. Although there are few data in the literature on disease transfer through the use of manure as a pond fertilizer, it does appear that the risk of disease tranismission via fish grown in such ponds is low. Further­more, such fish are nutritionally and economically beneficial for farmers and consumers. Fish are not susceptible to most infections of warm blooded animals (livestock and man); they are healthy and demonstrate good growth in well managed manured ponds. The main danger lies in the passive transfer of pathogens, e.g., Salmonella but there is a rapid attenuation of enteric microorganisms in nanured ponds in the tropics, probably due to high temperature, pH and dissolved oxygen. Fish raised in manured ponds should be washed and cooked well prior to consumption as a final safeguard.

The construction of fishponds may provide breeding sites for insect vectors of disease., particularly mosquitoes that may transmit malaria. However, mosquito breeding in ponds can be largely controlled by good design and management, in particular by preventing vegetation either hanging into or emerging through the surface of the pond (Feachern et al. 1983). Furthermore, the fish themselves may aid mosquito control by the consumption of larvae.

A far greater threat to public health in certain parts of the world is schistosomiasis (bilharzia), an occupational hazard to people who enter fishponds. The disease is caused by Schisrosoina,a helminth parasite for which the intermediate host is an aquatic snail. Schistosomiasis is a major insanitary disease of man and has increased with the construction of reservoirs and irrigation schemes. Although there are only small foci of infection in Asia, it is most widespread in Africa and northeast South America. Planned epidemiological and ecological studies must be carried out before the implementation of water development schemes in the tropics, including fishponds. A carefully defined package of chemotherapy, health education, sanitation and snail control is required to control the disease (WHO 1980). For most water-borne diseases, the environmental impact of fishpond development is broadly analogous to that of irrigation development, which has been recently reviewed by Verma (1986).

A recent report on the risk of human influenza pandemics from close association of pigs and poultry on Asian farms singled out integrated livestock-fish farming as being a potential source of increased pandemics, thereby creating a negative impression of the acceptability of integrated agriculture-aquaculture farming systems in gencial (Scholtissek and Naylor 1988). They suggested that pigs may be "mixing vessels" in which normally separate avian and human influenza virus reservoirs meet, leading to genetic reassortment and the origin of new human pandemic influenza strains. The promotion of integrated systems in the third world should not create potential human health hazards. However, the inferred link between aquaculture development and human influenza was grossly overstated. Pigs and poultry have been brought together withoutfish on traditional farms in Asia and Europe for centuries. Furthermore, integrated agriculture-aquaculture tarming systems involving pigs, poultry and fish are rare and likely to remain so. Such a special case should not hamper the development and expansion of beneficial integrated crop-livestock-fish farming systems. More-recent trends in livestock development in both East and West are towards monoculture because of management and marketing considerations and this also apply to the integration of livestock with fish. These points were made in subsequent correspondence (Edwards et al. 1988; Naylor and Scholtissek 1988).

Education A Systems Approach to Agricultural Development

It is essential to appreciate the concepts that underpin the application of systems thinking toagricultural development before a framework for education to advance the development ofintegrated farming systems can be constructed. The term education is used here to encompass alllevels of 'teaching about agriculture' from extension to farmers to tertiary level studies atuniversities (FAO 1984).

A systems approach to any activity starts with the concept that everything is connected and achange introduced in one part of the system will induce a change in other parts of the system.Whether this change will cause an unimporiant ripple or an irreversible wave is often difficult todetermine, particularly in complex systems involving the integrated farming of crops, livestockand fish. This implies that the study of parts of a system in isolation will not be adequate tounderstapj. the complete system or to solve problems that stand in th way of its design,constri:ction, repair or improvement.

Modern concepts and techniques for a systems approach were developed by militaryscientists out of a need to explore the total implications of alternative strategies to achievespecified goals. The value of the approach was soon appreciated ;n other fields and thetechniques have now found their way into many sections of science and industry under one name or another (Dent and Anderson 1971). Definitions and meanings of the words and phrases used are important first steps in systems work (Spedding 1979). So too is defining purpose. This iscritical because i. sets the framework in which discussion takes place and drives the decisionmaking process in systems operation.

The fact that there is something that can be called "a systems approach" implies that there must be:

- A philosophicalfoundation from which it derives (Popper 1959; Checkland 1984);- A body of theory upon which it rests (Boulding 1956; von Bertalanffy 1968; Campbell

1985); - A set ojprinciples to guide action, the first of which is to identify,, classify and describe

the systems in which one is interested in order to establish their initial state (Spedding1979);

- A way of proceeding. After it has bee:, decided what system is being considered and howit operates now, if the purpose is to improve it, the prime problem that stands in the wayof achieving this improvement must be identified and clearly defined. This is no easytask. In fact, Einstein once said that the definition of the problem is more important than the solution.

The next steps are: - analyze the problem in relation to the purpose of the system; - hypothesize a solution; - synthesize the system under investigation; - test the solution in the context of the system.

It is possible to proceed in one or more of three ways: - Accept the hypothesis as a reasonable estimate of the truth and go ahead and test it in an

ad hoc way;

31

32

- Test the hypothesis physically in a controlled, scientific way; - Test the hypothesis in an abstract way by using computer models in which changes in

systems variables can be manipulated. If the solution is not acceptable to those operating the system then the whole process must

be repeated.

A systems approach is highly applicable to the activity of farming systems research and development. It is an extension of a scientific approach (some say a mirror image) which will make traditional studies of agriculture and aquaculture more rewarding. Furthermore, it is evolving as a way of bridging the gap between the generation of knowledge by research and the use of that knowledge to improve the output of products and money from farming systems. The computer may be one of the tools it uses in addition to the backs of envelopes. Modelling and the construction of diagrams may be important techniques to employ while the collection and analysis of data will usually be an essential first step to establish the nature of the system under investigation.

The important point emerging from all these efforts to come to grips with the real forces that underpin agricultural development, and they are not always technical, is that more and more people are beginning to wonder what the world really looks like from a farmer's point of view; in effect itis becoming respectable to stand "in the shoes of a farmer" to find out what his purposes are in order to be able to identify opportunities that could he available to him, to define the problems that block their achievement and to seek acceptable solutions. We need to be able to formulate these problems in such a way that solutions to them are testable before a farmer commits to what might be an inappropriate course of action.

EducationPrograms

Recent reviews of fisheries education needs and opportunities have been prepared by Chua (1987) and ICLARM (1986). Although education in these publications is considered at all levels, only tertiary education at post-graduate level is dealt with in the present study because this is the level at which research and education are most interdependent and at which education is most urgently needed - to educate future educators. Universities and similar institutions are often thought of as centers of teaching and research However, they can also be thought of more simply as centers of learning. Teachers, researchers and students are all involved in a learning process.

Tertiary education programs in agricultural/farming systems vary according to country and the circumstances of the institutions involved. However, the purpose of programs based on a systems approach is to produce people who are intelligent rather than informed; primarily biologists, but who are unafraid of either economics or mathematics or getting their hands dirty and who are psychologists and diplomats as well. Such people will be opportunity makers and takers and problem-solvers.

A criticism of a systems approach is exemplified in the statement "systems people know a little bit about everything and not much about anything". This criticism should not be taken any more seriously than the criticism of traditional subject specialists who are said to "know more and more about less and less". Both types of people are necessary; it is how their knowledge is used that is important. Furthermore, the knowledge base for agricultural production is now so great that the mere manipulation of the margin between the costs of inputs and the prices for outputs can produce either huge surpluses of agricultural products (as exemplified by the Common Agricultural Policy of the EEC) or tragic deficits as exemplified by the socially disturbed and overly controlled economies of some countries. What this really means is that the output of agricultural products is powerfuliy influenced by the rewards farmers receive for he efforts they make and the risks they take. This simple fact is sobering. Professional scientists must keep their feet on the ground and not become unduly impressed with the technical advances that they can achieve. The potential impact of new technology on farming systems is dependent upon a wide range of social and economic factor- (Fig. 12).

33

/ International Trade in Agricutua

Politicians Economists Intra-national Trade Traders

Agricultural Systems

Farmers Agriculturalists Farming Systems Traders

Subsystems

Plant and Animal Components Scientists Cmoet

Fig. 12. Levcls of focus in agricultural dcvelopment (from Gartner 1984). Factors in the external environment of a fanningsystem (exogenous factors) niay show greater potential for improverent than factors in the internal environment of afarming system (endogenous factors). The term farming systems may then inhibit the teaching and understanding of thewider picture. To overcome this the term agricultural systems is used to encompass the delivery systems for providingessential naterials and services to farmers and for getting products to consumers. The question mark indicates further possible levels of focus in this hierarchy of systems.

A systems approach to agricultural development provides a way of examining changes inthe components of an agricultural system that will reveal their effects on the system as a whole.Students who develop knowledge, skills and attitudes in this direction should be able to assist inunravelling some of the complex agricultural issues of their time. How this might be achieved can be divided naturally into three teaching activities:

- Teaching what we know (KNOWLEDGE); - Teaching how to discover what we need to know (RESEARCH); and - Teaching how to combine both to make improvements in identified agricultural/farming

systems (DEVELOPMENT).Traditionally, teaching what we know provides the bulk of an undergraduate program; it is

mostly receptive learning, perhaps with a short-term rcsearch project. Teaching how to discoverwhat we need to know is normally confined to post-graduate programs at M.Sc. and Ph.D. levelsusing the classical methods of science. Teaching how to combine both to make improvements inidentified agricultural/farming systems has only begun recently: experiential learning is usuallydominant, even at the undergraduate level (Bawden et al. 1984), with receptive learning on theperiphery of the education process which takes place within a systems context. Courses on "tools and techniques" are taken as required.

34

What constitutes an improvement is, of course, always open to debate, which relates back to the all important question "from whose point of view?" - government, researcher, farmer, trader, consumer? The likelihood of improvements being made can often be traced back to who holds the balance of power amongst these and other participants in the operation of an agricultural system. Therefore, an educational program should be heavily oriented towards practice so that the students at some time during their training "stand in the shoes" of a farmer, an extension worker, a researcher, a trader, a banker and a policymaker.

The Needfor Researchin Association with TertiaryEducation

An active, visible research program is an essential component of the overall education process in a systems approach to agricultural development. Highly focu.d student thesis research on problems involved in the improvement of identified farming systems can certainly contribute to knowledge, but students come and go. A professional research program on the opportunities and problems within an institution's area of influence is needed to complete the structure of an educational program. Student research then takes place in the context of an on­going program, not in isolation, but in the greatest tradition of learning with the student working with and beside the teacher. Furthermore, the cooperation of farmers in this activity is essential. Without an understanding of their point of view, and continuity of contact, they are not likely to be interested in cooperating in the teaching program or in applying the results of the research work.

An Example of a Systems Approach to Educationin IntegratedFarming

Background and relation to national programs and other institutions

The work being done at the Asian Institte of Technology (AIT) is used as an example for the development of post-graduate programs in agricultural/farming systems. Full details are given in AIT (1988). This work is part of a larger regional program in education to develop systems thinking in key universities and institutes, the purpose of whicl is to link centers of learning arid research to the centers of production. Students come from professional and farming communities, drawing upon each other's knowledge and skills in interlocking activities. The program, initially organized as a UNDP/FAO Regional Project, established post-graduate Diploma and M.Sc. courses in Farming Systems at the following universities: Khon Kaen University, Khon Kaen, Thai!" ad; Zhejiang Agricultural University, Hangzhou, China; University of Peradeniya, Kandy, Sri Lanka; University of the Philippines at Los Bafios, Los Bafios, Philippines; Institut Pertanian Bogor, Bogor, Indonesia; National Institute of Agricultural Science, Hanoi, Vietnam. The national institutions were faced with the same questions that faced AIT: should programs have systems thinking and methods as the core of their activities with subject matter or the periphery or should traditional subjects and disciplines in agricultural/aquaculture science be supplemented by course work on a systems approach to agricultural development?

These questions have led to national programs that vary in style, content and delivery. Khon Kaen University, for example, has a fanning systems research group. Members of this group contribute to teaching farming systems as a subject at undergraduate and post-graduate levels. The post-graduate Diploma program includes farming systerns as core subject matter for the first time.

All programs sha'e the common themes of trying to present subject knowledge in a systems context and to involve farmers (as well as other participants in agricultural systens) very early on as partners in the educational process.

35

Other tertiary educational institutions in the Asia-Pacific region have established substantiveprograms in Agricultural/Farming Systems, each with their own style and emphasis. Amongthese are: Hawkesbury Agricultural College, Richmon- Australia; Chiangmai University,Chiangrnai, Thailand; and the South China Agricultural University, Guangzhou, China.A key difference between AlT and the national institutions is that the latter can zero in onnational issues relating to Farming Systems Development. However, AIT plays a vitalcomplementary role with its intenmational program by bringing together students from differentcountries. This drives home the point that agricultural development issues at the internationallevel can dramatically affect those at the national level (Fig. 12).

The M.Sc. program in agricultural ,ystems at AIT

Generalprinciples

The principic uf delivery at AIT is based on the Chinese proverb:

What we hear, we forget.What we sec, we remember. What we do, we understand.

Any debate on improving the efficicncy of small-scale farming systems in Asia should alsoinclude fish, the major traditional source of animal protein in many areas. Integrating fish with crop and livestock production adds to the complexity of the competition for the resources oflabor and capital but opens up considerable possibilities for increasing the output of focad andcash from the resources of land and water under a faimer's control. Therefore, practical work infanning systems development is focused on integrated farning systems involving crops,livestock and fish. The interrelationships in time and space between the components and resources of these farming systems are highly complex. Furthermore, the program concentrates on rainfed farming systms because these are widespread in Asia and have been neglected indevelopment studies. It is believed that if students can be taught within the framework of thesesystems, they S;hould develop the confidence to tackle any problem situation in agricultural oraquaculture development with imagination and ingenuity.

Structure

In setting up a study program to achieve these purposes, one is faced with existinginstitutional constraints. An exception to this is where one begins at the outset with a "Statementof Intent" to establish an institution based on a systems approach, such as was done for theInternational Livestock Centre for Africa (ILCA 1980).Within an existing institution the options are: a) create a ievolution, which can have negative consequences;b) slowly merge the new program with existin 5programs until an identifiable whole emerges, composed of teaching in terms of systems, components and techniques.The choice depends on circumstances and people. AIT is proceeding with the latter option.The major problem faced, which is typical of systems work, is what to include in the programand what to leave out, in order to fit within the constraints of the academic requirements of theinstitution. These include: time allocated to complete the M.Sc. degree program; residentini

recuirements; the minimum number of units required in terms of formal lectures and practicalclasses and thesis regulations.

Practical classes are most demanding but they are useful for introducing material that cannotbe presented in lectures; in fact they generate a demand for relevant knowledge and informationwhich assists with the choice of material to be included in lectures.

36

The AIT Masters program specifies a total of five terms (semesters), each of about 12 weeks, spread over 20 months. A minimum of 30 units of lectures and practicals are taken over three terms and a research thesis for 25 units of credit must be presented. A high staff student ratio is required because of the number of contact hours demanded by practicals: one full time professional (with appropriate technical assistance back-up) to six students. Three full time professionals with systems training are required to deliver an effective program in association with other faculty.

The core of the program consists of lectures and practicals on agricultural systems. The lectures:

- introduce systems concepts which provide the framework wherein course material delivered on soil, water, plants, animals, men, money, machines and markets can be applied in real life situations;

- expose the methodology of farming systems research and development; - look at the wider issues involved in farming systems development in order to balance

attention to the farmer on his farm with attention to government policy, market forces, input supply and consumer habits;

- establish an understanding of the need to reconcile many different points of view in the process of decisionmaking for agricultural development.

The practicals: - give students experience in the recognition of opportunities for improving agricultural

systems, in the identification of problems that stand in the way of their realization and in the finding of solutions to those problems;

- involve students in the development and continued improvement of an on-campus teaching farm of 2.5 ha involving the integration of crops (rice, maize, fruit and vegetables), livestock (buffalo) and fish (tilapia and carps);

- associate students with component research on an area of 2.5 ha adjoining the teachingfarm on problems identified during its design, construction and operation;

- take students off-campus onto farms and into industry and government to investigateproduction problems, the supply of inputs, the marketing of outputs, the availability of credit and the formulation of policy.

Additional courses are given on: Crop Production Systems; Livestock Production Systems and Aquaculture Systems. These involve the essential biology involved in the "breeding,feeding, health and husbandry" of crops, livestock and fish, and their temporal and spatialrequirements, all set within a systems context. The link to capital and labor is provided byanother course, Farm Management Economics. The final recommended course is IntegratedFarming and Waste Recycling in order to tie the above courses together and establish the links and energy flows in the food chain.

Elective courses to satisfy degree requirements can be taken in related subjects, such as agricultural engineering, rural development and computer science.

Importantareasof concern

The program pays attention to two areas of major ;-ncern - water availability and pre- and post-harvest losses.

Rainfed farming is of great importance throughout the third world. Water is the key resource and this is brought home to students in the operation of the integrated farm under rainfed conditions. No irrigation water is allowed. Therefore, topics such as rainfall probabilities, soil moisture management and on-farm water harvesting, storage and distribution are of greatimportance and are highlighted in the practicals. Water is of course essential for fish. Here attention is given to its quality as well as its availability throughout the cycle of the seasons.

37 Efficiency in agricultural systems can be improved by increasing absolute output per unit ofsome resource, but a most neglected area for achieving it is in the reduction of losses of what hasbeen produced. A substantive way to bring students' attention to this as well as to theenhancement of product marketability is considered essential.

Some problems

The educational process in agricultural systems requires working at "right angles" toconventional subject knowledge and research. Instead of attempting to reduce .reas of ignorancein a particular subject by delving more deeply into it, it is necessary to probe availableknowledge in many subjects and disciplines for the facts required to develop a view of a wholeagricultural system (Fig. 13). This view needs to be focused at a sufficient level of resolution(Fig. 12) to determine the consequences of different policies and decisions that affect productionand profit and other system properties.

Economics

Social Agriculture Sciences

~Biology

Fig. 13. Diagrammatic representation of agriculture as P subject with some of theoverlapping disciplines involved (from Spedding 197)).

A problem with this type of work is that in attempting to build bridges among education,research, extension and practice, one can be uncomfortably isolated at times from the security ofthe foundations on which these recognized activities rest. Students may suffer this discomfortvery early on in their course work. They may feel unable to compete with other students insubjects that are new to them, which affects their self-esteem. However, t!ley slowly realize thatthey are becoming experts in their own right in viewing systems as a who-' rather than asseparate parts and that they can absorb relevant subject knowledge more effectively. Then comesan understanding of how the manifestation of a problem (opportunity) in a system is often a longway from the cause (stimulus). Finally comes the perception of system dynamics involving rates,levels and interactions.

38

A second problem confronting students is the difficulty of accepting the concept of a hierarchy of systems and that a component of one system could b a system itself with its own botundary and properties. Once this is accepted, component study arid research becomes just as valid and important as whole systems research; so long as the problem is generated by the system to which the solution will apply.

The third problem and the most difficult in systems work is "how to start?" The need at times for great feats of imagination and ingenuity was mentioned earlier, but some simple advice to students is "silart!" by asking the questions: what do I want? what do I have? what do I need?

Another problem common to international institutions like AIT is language. First, there is the difficulty of translating between cultures, particularly when it comes to abstract concepts. Second, there is the more practical difficulty of language when students carry out off-campus/on­farm field work in Thailand. This has necessitated additional Thai staff and places what must be construed as a helpful burden on Thai students. The problem disappears with the establishment of national education programs in agricultural/farming systems development, such as those that have been established through the regional program.

Employment Opportunities

How will the graduates of such educational programs with a systems perspective be employed? A logical place for a graduate v :th interdisciplinary training is in agriciltura extension because farmers must integrate their activities daily, seascnally and annually. However, some problems affecting the performance of farming systems are quite complex and may not be resolved in the day-to-day activities of an extension officer. Farming Systetms Institutes are being formed to tackle complex problems and to bridge the gap between subject­and commodity-oriented research and extension services to multicultural small-scale farming systems. Such institutes will require graduates with a systems perspective. However, it is still not enough just to produce people with a systems perspective; job titles, conditions of employment, promotional prospects and career ranges will have to be established. This does not preclude the wide variety of jobs in privat,. industry concerned with agriculture which are appropriate for a person trained as an opportunity maker and taker and as a problem solver.

Points of Vulnerability

The approach LO tertiary education in agricultural systems outlined here is somewhat revol,,tionary, especially as far as the integration of teaching and research on aquaculture within the context of agricultural systems is concerned. Aquaculture educators have yet to get involved in the mainstream of agricultural research and education. The approach, like all new approacheshas a number of points of vulnerability. These include:

A. What happens if the demand for tertiary educational programs in farming systems development outstrips the rate at which qualified, experienced teachers can be de-,eloped?

The conservative answer is to make haste slowly. It would be better to have a.few carefully thought-out programs - properly staffed, adequately funded and producing skilled people - than the opposite. Still, there is no better way to learn how to set up a program than "to have a go"; failure is often a very fast way of learning! In fact, succe., s is generaily the outcome of a long history of overcoming difficulties, if not failures.

B. Students may be discouraged when they re-enter the working environment if their superiors do not appreciate their newly acquired skills and fail to utilize them effectively.

39

C. Post-graduate students seeking to pursue theses in systems work may be penalized becausetheir work does not fit the existing academic requirements of sume institutions. Applicants forpost-graduate programs may be rejected because their background and qualifications do not fitthe requirements of traditional departments. These difficulties are especially likely in institutionswhich attempt to offer new programs in farming systems and which face a transitional periodfrom their former traditional approach.

D. It will be essential to have international and national lead centers of excellence from whichinstitutions new to a systems approach can draw support and guidance.

E. Specialists in agricultural/farming systems will need to meet regularly to discuss the way inwhich the subject matter is or should be developing; otherwise it will remain static andunresponsive to the changing needs of agriculture and aquaculture.

Despite these points of vulnerability, it is apparent already that this interdisciplinaryapproach to education for farming systems development will expan,' to involve institutions throughout the third world.

An Institutional Framework

General Considerations

Geograph ical/cli m at ic

Where can research and cducation for the development of integrated farming systems in the tropics and appropriatc subtropical regions best be conducted? The rational answer is undoubtedly in or close to the fairing areas where the results will be utilized or in a corrparable ell Vir )nicit rather rhan in artificial 'laboratory' conditions e!. , e re. Unfortunately, most rcscalch and teaching institIutions in the humid tropics and subtropics, where integrated fanning has losl prol-rIsc, require considerable strengthening to carry out successf, l programs. This a)lics palrticularly t thse in Africa. Therefoin the experience of the few strong institutions investigating the devlopment of integrated agriculttreaquacLlttrre systems Iocated in the trop ics is invaluable.

These institutions arc in Asit: AIT, ICLARM and its cooperators, and soic institutions within the AICI/AO Network of Aquaculture Centres in Asia (NACA). They constitute vital assets for the future development of tropical integrated farming and merit strong support to sustain their programs and to cxp'uinI their activities to help others. International cooperation between such established groups and emrc-gent groups in other regions across the tropical belt is likely to be iore art rctivet() (tICn(rs anld more valuable to the researchers concerned th.an isolated eftfrts or the more famili'ar North-South linkages. In particular, the concept of integrated farning research cooperation, educational linkages and technology transfer in a South-SouLth mode from Asia to Africa and other regions has great appeal. Some of the most proddrctive Asian aquaculture systems are based on African fish (tilapias) whereas the potential of tilapia culture in Africa remaius unrealized because of a variety of technical, socioeconomic and institutional constaints. This applies especially to smalj- and medium, scale aquaculture and integrated farming. The need for increased interregional cooperation amongst tropical institutions in research and education for the development of integrated farming is clear.

There is still of course an irmportant role for nontropical institutions, including developed­country universities, in such research and education. Basic labor'itory research - for example, on the physiology and biochemistry of pond biota --can be performed wherever good laboratory facil ities exist. Many of the more expensive and sophisticated items of analytical and measuring equipment needed for such studies are difficult and expensive to maintain in the humid tropics; it makes little sense to install them in third-world institutions that have chronic recurrent funding problems. However, there is no substitute for performing tropical farming research and education in the tropic.s iii projects involving the study of s.ystens that are interactive with the natural environmetat.This atpplies to Studies on fish (individuals, populations and communities), nutritional and environmental physiology, control of reproduction, parasites and diseases and above all to integrated farming,Systems that have crop and livestock subsystems in addition to fish, and to the highly locition-specific factors involved in certain aspects of social science analysis.

Institutions located in the subtropics can perforn useful work for application tinder their local conditions but cannot bv.e effective leaders of a program targeted mainly on the tropics. For example, in Israel there has been a highly productive research effort on the application to fishponds of livestock wastes supplemented with inorganic fertilizers and feeds. Israeli summer temperatures are tropical, but winter temperatures prevent the growth and reproduction of fish.

40

41

Israeli research has therefore been directed towards improving Israeli technology andmanagenent practices, iMcludifig overwintermig of tiapias with production cycles much longer:!, ,ould apply to tilapia culttre in tile tropics, intensive hatchery systems and largeproduction ponds, all under unique socioeconomic conditions. Similar considerations apply tothc resc; Fc h prog"rans 4 some teroperate f'uropcan cotutries.In iIe People's Republic of' China, the 'ancestral home' of integrated farming, the climate isnot like thosc of tropical third-world countries. The ADCP/NACA Regional Lead Centre inChiln: is at ithe I re.sawater Fisheries Centre, WuXi, Jiangsu, where winter temperatUres fall to5 (. This is hlow the lower thermal tolerance limit for tilapias, which must therefore be held in Zgren huse s from Novembcr to May (FA() 1983). The species that survive there. principallycatrps, ti ive Iinited appeal in many .ntner countrie: (Pu!lin 1986). No systems can be studied atW\x i whiJh rCquti.-Ca tHinin tcrrtited growth phase of more than eight montis. This centerconecenrtesonm integrated tart r a rcsctare h and educalion on Chinese systems. It is not well­sited, ht wcvcr, to ;lay Ima 1,*0 inl a program Focused on the tiopics. The same applies to the,\)( 'I' intCrregina1 ceCeric tI1)uarv Research and education for tile development of tropicalat naIaClltnrca he c Ij erIrtC.""IiycMt.shea' ri tsile the tropics. Considerable crotributionshave beet 11,idc h ' ('hLnese, Fauropean and IsraelIi researchers to the advancement of aq(luaculturebut systcil>, !Cvc1( pcd in ihesc ,arcas and the research data that support them cannot betransferred directlyV .o tropical thlr -world co utries. Experts from nontropical institutions cantherefore hest pMtrticippe in future programs by contributing their knowledge to activities located

It mtke: no sense to separate the study and implementation of inland aquaculture fromagriculture in third world cotintries. Crop and livestock farmers will be the fish farners of thefuture. The chalcncI is !0 integrate aquaculture into existing farming systems as a profitablesubsystem, there Khinip)mv ing the productivitV and profitability of farms.To accoilaplish the rc war(h and educational activities outlined in this framework willreqree the c( peMratit t ,Iresearchers from different disciplines - aquaculturists, agronomists,hi(h(gisis (principal lish hI io is, microbial ecologists and other specialists on pondbi~oia engineers, aitrming s\ ,ltcnu,spccialists, livestock specialists, economists and other socialscientists. It cannot hc done hv ihlogically oriented atluaculturists alone. Thus, a program that is truly itrrdisciplinar' is FecCdrd. Aquaculttire research must be brought into the mainstream ofagricUIt nra resca rc ht. '1hiP req nires a new approach because aquaculture research to date hasbeen large .y the province of fish biologists. They and their donors have rarely recognized thataquact lIttar must he seen in the broader context of other food producing systems, principallyagrieCuIt Lire and capture fisheries. Indeed, aqtnacLl ttire in this context will use many of the same resotees and nmarketing chananels.

A new research initiative is reuired. A twofold program is desirable: genetic improvementof appropriate cultured specie:- and interdisciplinary (biotechnical and socioeconomic) researchto i.mprove culture system (ICIAARM 1988, Fig. 14). These two components must be interactiveCiplementary aind I )CtLis on small- and medium-scale farms. This concept may bechallenged by those who prefer separate discipline-specific programs in disease- engineering,

,r )ii

nutrition, reproduction, economics and sociology. However, the advantage of a simpler twofoldframework is that research expertise in these and other disciplines can be co-opted to assist themain research thrust ,as ant when required.Ti lapia has the widest acceptance and besi prospects for international programs (Pullin1985). The Second International Sympositma on Tilapia in Aquaculture. held in Bangkok, 16-20March 1987 drew 258 participants from 40 countries (Pullin e al., in press). The largestscientific sessions were on genetics/reproduotion and culture systems/management. Carps arealso important, particularly in some Asian countries, e.g., Bangladesh, China, India, Indonesia,Pakistan, Nepal, but carp culture has far less scope for growth worldwide than tilapia culture

because of market acceptance problems (bony flesh and poor keeping qualities) and therelavive!y sophisticated hatchery technology required for some species (Pullin 1986). Tilapia is

----

42

Aquaculture

Coldwater Temperate]Tropical(I+Subt.ropical)'

Coastal Open Sea Inland:

Export Markets Mixed Markets

Domestic Markets'

Large Large Integrated Fish Farms F Small (+ Medium) Farms

Integrated Farms 4

Crustaceans Others S Finfish'

Caritivores Specialized FeedersSHe rbivores/Dot ritivores/1

Omnivores"

Catfishes Others

I I Carps)'

Diseases Reproduction Nutrition Economics Socioeconomics Engineering

\-S rovmen ymp

Improvementar

Research Focus'

Fig. 14. Focus for a proposed new initiative in aquaculture research. 'I.e central boxed categories represent the choice of research; hemes. The dotted Iies indicate inputs from research in other disciplines as and when required.

Explanatory notes: 1. Climate. Aquaculture has most scope for growth in the humid tropics (and parts of the subtropics). 2. Sector. Inland aquaculture has the most potential. Open sea and coastal aquaculture have greater environmental constraints. Farmers are better suited to fish husbandry than fishermen. 3. Markets. Exports cam f:reign exchaage but markets may be short-lived and benefits limited to the wealthy. By concentrating on domestic markets, rural farmers can improve their own livelihood and produce fish at prices affordable by the rural and urban poor. 4. Farming system/scale. Aquaculture is most attractive as a subsystem, integrated into small- (and medium-) scale farming systems. Such integration can be applied in rainfed and irrigated systems. 5. Target group. The finfish offer better prospects for s:;tainable livelihood improvement than the crustaceans and other groups. 6. Feeding habits. llerhivorous/detritivorous/omnivorous fishes are better for culture in integrated farming systems than camivores or specialized feed~ers. Crop byproducts and livestock excreta can be used as fish feeds and pond fertilizers. 7. Speciesfocus. The tilapias (and carps) are the best species to use. The catfishes (some of which are carnivorous) are less suited to integrated farming. They require intensive feeding. 8. Research focus. Genetic improvement of tropical fusfish can make a major impact on fish production similar to those achieved for crops and livestock. It has not yet been attempted through a well-focused program. It must be interactive with culture systems improvement research, concentrated on integrated agriculture-aquaculture farming systems.

43 probably the easiest fish in the world to breed and to grow in a wide range of systems. Therehave been market swingi towards tilapia and away from milkfish in the Philippines and Taiwan.Even in China and the Indian subcontinent, which have traditionally preferred carps, interest intilapia culture is increasing rapidly although experience is very limited. However, carps areincluded in the research focus defined here because some can be grown in productive polyculturewith tilapia and have greater tolerance to the seasonally cool temperatures of the subtropics.

Institutional

Unfortunately, many existing aquaculture research facilities in the tropics and subtropics arebadly sited and/or poorly supported. To lead research and education activities, an institutionmust fulfill similar criteria to those that have ensured the success of the international agriculturalresearch centers: availability of land and water, good communications, schooling, housing,transportation links and security. Without these ii is exceedingly difficult to attract and retain theservices of high quality staff and to conduct sustained research and educational programs.A core prograin is essential to provide strong leadership and coordination for the researchand educational framework defined here. This must be a program of active research andeducation, not just a secretariat and information base.

The cor, program should be independent to insulate it from the frequent shifts in objectives,policy and funding support that are so characteristic of navional agencies and governments andwhich consequently affect the programs of institutions dependent upon their recurrent support. A core program needs :ustained objectives and sustained funding.

The involvement of international, regional and national institutions and researchers outsidethe core prnogram should be sought by means of a network, led and coordinated from the core.This is how interregional cooperation can best be achieved. Research advances and technologydevelopnment by the core and by relatively strong Asian institutions participating in the network can thereby be shared with institutions in Africa and other regions.The benefits and cost-effectiveness , networks are described by Plucknett and Smith (1984,1986). They point out that many of the .enatinal Agricultural Research Centers (IARCs)contract o, t basic research to "avoid duplication of effort and to keep in touch with upstreamdevelopments". They give as an example the cooperation of botanists, taxonomists, cytologists,geneticists, ecologists, biochemists and plant hreeders" in crop development research. Itobviously makes sense t,-) tap the exl'eiise of strong university departments and researchinstitutions from all over the world I r s;uch interdisciplinary work. The same applies to researchand education in integrated farming. 1-lore the expertise and participation of the crop andlivestock IARCs and othe, centers (such as Winrock International for livestock) would be of great value.

ICLARM coordinates the Asian F~isheries Social Science Research Network (AFSSRN): anetwork of institutions, the work of which encompasses aquaculture and fisheries (Maclean andDizon 1987). The AFSSRN is planning increasing activities in integrated agriculture-aquaculturein Southea.": Asia and thus could potentially provide some of the social science expertiserequired.

A list of some institutions with ongoing or poten::, ' interests in integrated fanning (i.e.,possible participants in a global network) is given in Table 4. This is not an exhaustive list andcould be supplemented with many more institutions, particularly agricultural institutions.However, the temptation to involve from the outset as many institutions as possible in anetwork should be avoided. It may create a good political impression but has pitfalls. The effortsrequired for liaison activities for large complex i-stitutional networks aad the thin spreading offunds can prejudice the goals of making rapid research and educational advances. Therefore, it isargued here i:at it is better to concentrate on a well defined core program and a network of asmall number of strong institutions with proven ability to advance research and educationalgoals. It would be the responsibility of each of these institutions to expand the network further.There is also scope for development of a larger network of individual researchers, the main

44

Table 4. Some institutions with current or potential interest in research and education for the development of integrated agriculture-aquaculture systems. This is not an exhaustive list and there are many other institutions that can contribute to integrated farming research and education, particularly if the sectoral barriers between agriculture and aquaculture are lowered.

A. Suggustwd Core Program and Network Coordination

Asian Institute of Technology, Bangkok, Thailand International Center for Living Aquatic Resources Management, Manila, Philippines

B. Potential Network Participants in Asia, Africa and Some of the Developed Countries

1. ,4sia

Bangladesh Bangladesh Agricultural Pasearch Council, Bangladesh Agricultural University, Mymensingh; Fisheries Research Institute, Mymensingh

China South China Agricultural University, Guar.gzshou; Pearl River Fisheries Institute, Guangzshou; Freshwater Fisheries Centre (NACA Regional Lead Centre), Wuxi, Jiangsu; Zhejiang Agricultural University, Hangzhou. Zhejiang

India International Crops Research Institute for the Semi-Arid Tropics, Hy.derabad; Central Institute of Fresh­water Aquaculture (NACA Regional Lead Centre), Dhauli, Bhubaneshvsar; Tamil Nadu Agricultural Univer­sity, Coimbatore

Indonesia Agency for Agricultural Research and Development; Directorate General of Fisheries; Institut Pertanian Bogor, Bogor

Malaysia Universiti Pertanian Malaysia, Serdang, Selangor

Nepal National Aquaculture Centre for Training arid Allied Research, Janakpur and Integrated Farming Lead Station, Hetauda

Philippines Central Luzon State University, Munoz; International Rice Research Institute, Los Ranaos and its Asian Rice Faiming Systein Network; University of the Philippines at Los Banios; University of the Philippines in the Visayas

Taiwan Asian Vegetable Research and Development Center, Tainan

Thailand Khon Kaen University; National Inland Fisheries Institute (NACA Regional Lead Centre), Bangkok

Sri Lanka University of Peradeniya; Ruhuna University, Matara

2. Africa

Cameroon Institut di Recherches Zootechniques, Yaounde; FAO/UNDP projects

Ethiopia International Livestock Centre for Africa, Addis Ababa

Ghana Institute of Aquatic Biology, Achimota, Accra

Cote d'lvoire Centre de Recherche; Oc;anologilues, Abidjan; Institut des Savanes, Bouake; West African Rice Develor. ment Association, Bouak6

Malavvi ICLARM/Malavv! Department of Fisheries, Lilongwe; Bunda College of Agriculture, Lilongwe and Chat­cellor College, Zooiba; University of Malawi, International Crops Research Institute for the Semi-Arid Tropics/Chitedze Agricultulal College, Chitedze

Nigeria International Institute for Tropical Agriculture, Ibadan; University of Lagos; Rivers State University, University of Calabar

Rwanda Universite Nacional, Rwanda

Zambia Department of Fisheries/FAO/UNDP Project, Chilanga

Zimbabwe Department of Agriculture, University of Zimbabwe, Harare

Continued

45

Table 4. (continued)

3. Developed Countries

Australia Department of Primary Industries, Brisbane; -lawkesbury Agricultural College, Richmond

France Centre Technique Forestier Tropical, Nogent-sur-Marne, Ecole Normal Superieure Agronomique, Tou­louse; Instit it National de la Recherche Agronomique, Jouy-en-Josas

Israel Agricultural Research Organisation, Bet Dagan and its Fish Culture Station, Dor, Hof Hacarmel

Jap-an Asian Productivity Orgarisation, Tokyo; United Nations University, Tokyo

Netherlands Agricultural University, Wageningen

Norway Agricultural University, As.

United Kingdom Institute of Aquaculture, University of Stirling, Stirling; University of Reading, Reading

United Statesof Armerica Auburn University, Auburn, Alabama; Oregon State University and associated US universities within theConsortium for International Fisheries and Aquaculture Development; Winrock International

C. Potential Future Network Participants in the Caribbean and Latin America

1. Caribbeat

Jamaica AquJculture Progran Department nf Zoology, University of the West Indies, Kingston

Puerto Rico Umiversity of Puerto Rico, Mayaguez

US VirginIslands AC.aculure Program, College of the Virgin Islands, Kingshill, St. Croix

2. Latin America

Coiornbia Centro Internaconal de Agricultura Tropical, Cali

Mexico Instituto de ivestigaciones sobre Recursos Bioticos, Xalapa, Vera Cruz; Centro Internacional de Mejoria­inento Mai/ y Tri go, Mexico City

Panama Direccion Nacinnal (It?Acuicultura, Ministerio de Desarrollo Agropecuario, Panama City

Peru Universidad Nacional Agraria, La Molina, Lin i; Centro International de la Papa, Lima

functions of which are provision and exchange of information and results between members.ICL ARM, for example, operates a highly successful network of this type in fisheries science: theICLARM Network of Tropical Fisheries Scientists with about 700 members in 80 countries(Munro and Pauly 1982). ICLARM launched a sister Network of Tropical Aquaculture Scientists(NTAS) in mid-1987. The NTAS has integrated agriculture-aquaculture research as one of itsmajor themes (Pullin and Paguio 1987), and its membership currently exceeds 200 individuals.

On-farm activities

In tropical third-world countries, there. ..e many experimental aquaculture stations whichfunction poorly. Most were built with no clear objectives other than broad ideas to develop anddemonstrate aquaculture. Most lack realistic recurrent funding. A further common fault is thattheir ponds are often too few in number to permit adequate replication of treatments an('/or toolarge to manage adequately. The most valuable on-station research results in tropical Asia havecome from ponds in the range of 200 to 1,000 M2 , particularly 200 to 400 m2.

46

There are also large, so-called demonstration facilities. In reality, these have little to demonstrate other than attempts to increase their on-site production year-by-year by what is really guesswork: changing various inputs simultaneously, e.g., species combinations, stocking densities and nianagement practices. This has been a feature of much on-station 'research' in both Africa and Asia and is still widely practised. The fish yields achieved may impress politicians and funding agencies but rarely give insight, into the underlying basis of aquaculture prodUction in a form relevant to small-scale farmers wro are seeking to adopt or improve systems based on their own limited resources. Moreove,-, the financial analysis of on­station/demonstration farms are usually Very special cases even though it is often presented as representative of industry economics which it is not.

Some researchers have worked successfully with farner cooperators but the concept of investigative on-farm research in integrated farming (involving cooperation between farmers, researchers and extension workers from conceptualization through experimentation to analysis, publication, dissemination and implementation of result,;) remains poorly developed. Where there are potential farmer cooperators, and particularly where these have close working relationships with extension services, such activities can generate important data and direct benefits to farmers. Moreover, the compilation of databases from working farms can give valuable insights into the most inportant factors affecting the productivity of a given system and can be a valuable addition to databases obtained from on-station reseatc<h. This is of course only possible where a sign.ficant number of fanner cooperators can be easily reac,,.z _"nd are willing to cooperate with researchers and extension worLers.

Johnson and Claar (1986) argue for stronger linkages between farmers and researchers to emphasize research under farm conditions. They report that farming systems research in Zambi,t and its relationship with extension have been improved by the creation of posts for Research Extension Liaison Officers. This facilitates the development of a research-extension continuum. Phiri (1986) has surveyed the "instituional environment" for agricultural development in Mala 'i and concludes that there is a lack of coordination bctween researchers from different institutions, extension workers, planners and policynmakers. He argues strongly for "multi-disciplinary on­farm research coupled with 'bottom-up' (rather than 'top down') planning ... ". Lightfoot (1987) stated that farmer participation in farming systems research is vital. He reviews 'indigenous research' b1farmers and confirms that many farmers understand well the concepts of experimentation and controlled input-output trials. However, he also points out that such indigenous research is slow; that farmers' knowledge is hard to elicit and that there are problems in implementing on-farm research that require the researcher to :;hare risk with the farmer and to replicate trials across farms for "quicker definitive answers". Despite these difficlties it is clear that "indigenous research" by farmers and formal experimentation can both generate important data.

It is concluded that on-farnr research, which is perhaps better termed 'adaptive field trials', can be exceedingly valuable, particularly in the tropics where on-station experimental pond and farm facilities are in short supply. It facilitates the generation of data and it.;use in technology development directly with farmer cooperators. Therefore, a program of research and education in integrated farming should involve on-fa:m trials and educational activities as well as on-station activities for two main reasons: (i) izwill increase the availability of experimental facilities and (given adequate safeguards and supervision) the flow of results, (ii) it will allow the testing of systems inder 'real world' conditions.

An Institutional Framework

A core program for research and education in tropical crop-livestock-fish integrated farming should undoubtedly be located in tropical Asia. Its potential for development in other regions (for example in Africa, see Preface) is less certain and leadership from Asia is vital. Then it can best serve the established and expanding integrated farming systems in this region through on­station and on-farm research and can educate students from other regions in the factors that are making Asian system.; successful. It should be located in the tropics where tilapias and some

47

carps can be bred and grown year-round. Moreover, it should be loca!ed in one or more countriesin which these groups are accepted as farmed fish. Only from such an environment can researchadvances and educational activities progress to serve a wide clientele. The clearest candidatecountries are the Philippines and Thailand. These countries have access to all the required fishspecies and practice a wide range of farming systems. ICLARM and AIT and their cooperatorsin the Philippines and Thailand are major institutions for integrated abriculture-aquacultureresearch and education in tropical Asia. The NACA Regional Lead Centre in India at Dhauli,Bhubaneshwar, focuses on the development of Indian major carp composite fish culture. TheNational Inland Fisheries Institute, Bangkok, pai :vf the Thai Department of Fisheries, is anothertropical NACA Regional Lead Centre. The NACA Lead Centre at Wuxi, China, is outside thetropics but could be a valuable network participant for research on Ciinese integrated farmingsystems. The Aquaculture Department of Southeast Asian Fisheries Developmerz Center(SEAFDEC), Philippines (which is also a NACA Regional Lead Centre), has not yet been amajor player in integrated farming research and education and does not yet have the facilities forsuch v.esearch. Moreover, as NACA changes from a UNDP project to an intergovernmentaloranization (NACA 1986), some of its member institutions in other Asian countries, e.g.,Bargladesh, Indonesia, Nepal and Sri Lanka, may increase their interests in integrated farming.

Turning beyond the co-', program and potential Asian network members to Africa and otherregions, here it is difficult to make firm recommendations. The African Regional AqitcultureCentre (ARAC), Nigeria (part of FAO/ADCP) has achieved little research progress in integratedfarming, but has included this topic in some of its training programs. The strong groups inintegrated farming involving fish subsystems in Africa tend to be project teams rather thaninstitutions. Good examples are a UNDP aquaculture project at Bouak6, C6te d'Ivoire (Nugent1987) and an FAO/UNDP project at Chilanga, Zambia (Gopalakrishnan 1987).The advent in 1986 of an ICLARM base in Malawi, with its initial activities tightly focused on an integrated farming research and educational project - Research for the Development ofTropical Aquaculture Technology Appropriate for Implementation in Rural Africa, funded bythe Deutsche Gesellschaft fur Technische Zusammenarbeit (GTZ), GmbH - should help tostrengthen institutional capabilities in southern Africa in the future. ICLARM has begun aprogram of interdisciplinary research on small-scale integrated farming in Malawi (in

cooperation with the Mala%,i Department of Fisheries and Chancellor College, University ofMala'vi) and has established linkages to help with orientation on tropical integrated farmingsystems for African researchers and culturists in Cameroun, C6te d'lvoire, Ghana, Rwanda andZimbabwe (Paguio 1987).

For Latin America and other regions (Caribbean and Pacific), the picture is similar. There are a few research and development group.; with strong interests in integrated farming,principally in Mexico, Panama and t'eru, but the overall capabilities of these regions inintegrated farming research and educatior, are limited. Moreover, the scope for growth offreshwater aquaculture in these regions is less certain than for Asia and Africa.Therefore, the institutioal framework recommended here is a strong core program at thecenter of a network of'a few strong institutions in Asia, with linkages to: 1) integrated farminggroups and institutions in Africa to foster Asia-Africa interregional cooperation and 2) linkagesto strong research groups worldwide for supportive research and edu'cational expertise(especially universities). Further linkages to Latin America and other developing regions arerecommended if strong interest, institutional capabilities and development potential can bedemonstrated. On a wider front, the ICLARM NTAS is expected to fulfill the task of linking andproviding information to individual researchers worldwide. Its members with integrated farminginterests will clearly benefit from contact with the institutional framework proposed here and canfeed in their own information to help research and education activities. The institutionalframework envisaged in general terms is depicted in Fig. 15. It is clearly necessary to survey thecapabilities and interests of poteniial network participants. However, lists of potential networkparticipants have been prepared (Table 4) based on information presented at theICLARM/UNDP Workshop, "Towards a Research Framework for Tropical IntegratedAgriculture-Aquaculture Farmning Systems", 15-17 October 1986 (see Appendix I).

48

Agriculturaland Aquaculture Institutions FAO/ADCP; IARCs;

Universities

F

AquacultureInstitutions raAgriculturaevingCore Program IARCs; Un iversities IARCs; NACA; ICLARM; AIT Preects li Universities t e tsh tFAg; inlateRMposLINDP;

N Latin America

linagso~tt t inDo r sevelopins

Latin America .0e

Agriculturaland

Aquaculture Institutions

IARCs; Universities

sducationfrg.15. A suggested frainwork for1 international program of cooperation iniesdarch and for the development ofintegrated agricuhlUe-aquacuhiure. Tihe ICI.ARM/AIT core piograrn has rcsponsibili!)y for ;csearch and edt- ational leadership

tohe selected according to ther expertise and interests.broad arrows (.oeplho) indicate the major network inlgesechicl 0hC cre,Africa and Asia. '11enPrFw arrows (-l -) indicate additional linka ges Cllie --- Fis-) possible future

and coordination ofa worldwide network of ins eitutions, llie

to tyer strong research and teaching groups. dotted arrows (-i ndicate linf o inSilionS t developdi g regi is comlre: tP, ooedinaion adkages ill ler st) A Aqiacurtre

c athdDeveloieni Programme;InstituteMT,Asian y of Tlchnology; FA, Fooog AgriCUure OrgItiiaion ofe Unied Nastans; BAraCs, .rie Center for Living Aquatic Resources Mainagem ni;mational Agricultural Research Centers; CI.ARM,]n.crilaliona]NACA Network of Aquaclihre Centlers in Asia; UNI)I', 1.;niled Natlions Development Progralmme.

Acknowledge ments

This study was made possible by a Preparatory A-,Fistance Grant (GLO/85/003) to -ACLARM from the United Nations Development Programme. Professor Peter Edwards is seconded to the Asian Institute of Technology by the Overseas Development Administration of the United Kingdom. Dr. Joseph A. Gartner was Chief Technical Adviser to the UNDP/FAO Regional Project RAS/81/044 Farming Systems Development in Asia: Crop-Livestock-Fish Integration in Rainfed Ar-as during the period over which this study was compiled but is currently seconded to the Asian Institute; of Technology by the Australian International Development Assistance Bureau.

References

AIT. 1986. Buffalo/fish and duck/fish integrated systerms for small .scale farners at the family level. AIT Research Report No. 198. Asian Institute of Technology, Bangkok, lhailand. 138 p.

AIT. 1988. AlT prospectus 1989. Asian institute of Technology, langkok, Thailand. 60 p.Blawden, R.J..IR). Macadoin, R. Packhoin arid 1. Valentine. 198.I. Systems thinking in the education of agricuhurists. I. Agric. Systems 13:

2(5-225.lcts, W.C. 1982. Multiple croppm and tropical lanning systeirs. Gower l ublishing Company Limited, Aldershot. 156 p.Bouling, K.E. 1956. Gieneral systenms thrry - the .keleton of science. Manage. Sci. 2(3).l3rewhaker, J.l. 1986. lcgurrinous trees :rid snirbs for Souiceast Asia and the South Pacific, p. 43-50. In G.J. Blair, D.A. Ivory and T.R. Evans(cds.) Forages ii Southeast Asian and South Pacific agriculture. ACIAR proceedings No. 12. Australian Centre for International

Agricultural Research, Cautx,-rr., Australia.Campil-l, R. 1985. Fishomilan's guidt: A systenis approach to c.-eativity and organization. New Science Library, Boston, Massachusetts, USA.Cairr, W.E.S. 1982. Chemical sfinnulatiit of lecding behavior, p. 259 -273. In T.J. Ilara (ed.) Cher.oreception in fishes. Developorents inAquacuihurC and Fisheries Science X.Elsevier Science Fublihers, Amsterdam, "lhe Netherlands.Checkl'mnd, P. 1984. Systems thinking, systeris practice. John Wiley and Sons, Chichester, UK.Chua, T.-. 1987. Aquaculturc training nceds in developing Asia. I),C Manuscript Report, ll)RC-MR 1,15e. Interational l)evcloproent Research

Ceitre, Ottawa, Canada. 51 p.Coclie, A.G. 1907. Fish cultivation in rice fields: a world-wide synthesis. Ilydrohiologia 30: 1-41.Colran, J.A. arid P. E-!wards. 1087. Fceding pathways and environmental csmstrvints in waste-fed aquacnltrire: hal:ace and optimizyation, p. 240­

281. In ).J.W. Moriarty arid !Y.S.V. Pullin (eds.) )ctritus and microbial ecology in aquaculture. ICLARM Conference Proceedings 14,420 p. Inenational Center for Lusvinj Aquatic Resources ManauRement, Manila, Philippines.

Costa-Pierct, B.A. and I).. Cra-en !987. FEstim.fing microbial production arid growth rates in aquaculture ponds rsing rates of RNA arid DNA synth'sis. Aquaculture 66: 69-78.

Cuenco, M. Mathe.'atical modelling oaf aquacultue systems. (Ir press)Cuencc, M.I., R.R. Ssi-kney and W. :. Grant. 1985.,. Fish bi,ouergetics and growth in aquaculture ponds: 1. Individual fish model developrent.

Iicol. Modelling 27: 169 190.Cucico, .I.., R.R. Ssickney and W.E. Grant. 19851). Fish bioncrictics and gowtl in aqiuactiture ponds: 11.Effects of interactions among si7e,

temperattre, dissolved oxygcr, ,,mionized ammonia arid food ott growth of individual fish. lEcol. Modelling 27: 191-206.Cucnco, M.L., R.R. Stickney arrid .. Grant. 195.Fih bioenergctics ard growth in aquaculture ponds. Ifi. Stocking size and feeding ra'e on fish productivity. icol. Mudelling 28: 73-95.

CRSP. 1986. Pond dynarnics/aquaculurc. CRSP data base; instructions to. !iara entry (edition 1:1). %,iril 1986. Pond I)ynamnics/AquacultureCollaborative Research Support Program, Oregon State University. 26 1'.

de la Cruz, C. and V. Cararigal, Editors. Proccedings of the hntemational Rice-Fish Fanning Research ard I)evelopment Workshop, Ubon,Thailand, 21-25 March 1988. International Center for Living Aquatic Resources Management, Manila, Philippines, and Intemationa;Rice Research Institute, Laguna, Philippines. (In press)

I)ent, J.1. ard J.R. Anderso,,. 1971. Systems managenentr arid agriculture, p. 3-14. In J.. l)ent and J.R. Anderson (eds.) Systems analysis in agricultural managetment. Joii Wiley and Sons, Sydney, Australia. 388 p.

Deveidia, C. 1985. Non-conventional feed resources in Asia and the Pacific. FAO/APIICA Publ. No. 6. 141) p. FAO Regional Office for Asia arid tie Pacific, Bangkok. 140 1).

Duckiarr, A.N. 1959. lihe current agricultural revolution. Geography 44: 71-78. l)Uckhar, A.N. 1966. Forty years on: agriculure irs restrospeCt and prospect. J.R. Agric. Soc. 127L: 7-16.I)uckham, A.N. and G.I. Masefied. 1971. Fanning systems of the world. Chalto and Windus, London, UK. 542 p.Edwards, P. 1980 A review of recycling organic wastes into fish, with emphasis on the tropics. Aquaculture 21(3): 261-279.Edwards, P. 1983. The future potential of integrated farming systems in Asia, p. 273-281. In 1.Takasi (ed.) proceedings of the Fifth World

Conference ott Aninial Production. Vol. 1. Japanese Society of Zootechnical Science, Tokyo, Japan.Edwards, P. 1987. Use of terrestrial vegetation arid aquatic macrophytes in aquacnlture, p. 311-335. i D.J.W. Moriarty and R.S.V. Pullin (eds.)l)etritus arid microbial ecology in aquaculture. ICLARM Conference Proceedings 14, 420 p. Intematioral Center for Living AquaticResources Managemrcnr, Manila, Philippines.

Edwards, P., C. Kwei Lin, D.J. Macintosh, K. l.eong Wee, I). Little and N.L. James-Taylor. 1988. Fish farming and aquaculture. Letters to Nature. Nature 333: 505-506.

Feachemn, R.G., D.J. Bradley, II. Garelick arid ).). Mara. 1983. Sanitation and disease: health aspects ofexcreta arid waste water management.John Wiley arid Sons, Chichester, UK.

FAO. 1981. Agriculture: towards 2000. FAO, Rome.FAO. 1983. Freshwater aquaculture development in China. Report of the FAO/UNDP Study tour organized for French-speaking African

countries, 22 April-20 May 1980. FAG Fish. Tech. Pap. 215. FAO, Rome. 124 P.FAO. 1984. Report of the FAO Expert Consultation on Improving the Efficiency of Small-Scale Livestock Production in Asia: A SystemsApproach. Bangkok, Thailnd. 6-10 December 1983. Report and Recommendations. FAO, Rome. 30 p.

49

50

Gartner, J.A., Editor. 1984. Proceedings of the FAO Expert Consultation on Improving the Efficiency of Small-Scale Livestock Production in Asia: A Systems Approach, 6-10 December 1983. Food and Agriculture Organization of the United Nations, Rome. 100 p.

Goldman, J.C. 1979. Outdoor algal mass cultures - II. Photosynthetic yield limitations. Water Res. 13: 119-136. Gopalakrishnan, V. 1987. Role of tilapia in integrated farming systems in Zambia. In R.S.V. Pullin, T. Bhitkasawan, K. Tonghuthai and J.L.

Maclean (eds.) The Second Intemational Symposium on Tilapia in Aquaculture. ICLARM Conference Proceedings 15. Department of Fisheries, Bangkok, Thailand, and International Center for Living Aquatic Resources Management, Manila, Philippines.

Grigg, D.B. 1974. The agricultural systems of the world: an evolutionary approach. Cambridge Geographic Studies 5. Cambridge University Press, Cambridge. 358 p.

Grigg, D.B. 1980. Population growth and agrarian change: a historical perspective. Cambridge Geographical Studies 13. Cambridge University Press, Cambridge. 340 p.

ttepher, B. 1978. Ecological aspects of warm-water fish pond management, p. 447-468. In S.D. Gerking (ed.) Ecology of freshwater fish production. Blackwell Scientific Publishers, Oxford, UK.

llickley, P. and R.G. Bailey. 1987. Food and feeding relationships of fish in the Sudd swamps. (River Nile, southern Sudan). J. Fish Biol. 30(2): 147-159.

Hopkins, K.D. and E.M. Cnz. 1982. The ICLARM-CLSU integrated animal-fish farm.,, project: final report. ICLARM Technical Reports 5, 96 p. International Center for Living Aquatic Resources Management, Manila and the Freshwater Aquaculture Center, Central Luzon State University, Mufioz, Nueva Ecija, Philippines.

Iloque, M.Z. 1984. Cropping systems in Asia. On-farm research and management. International Rice Research Institute, Los Bafios, L..aguna, Philippines. 196 p.

ICLARM. 1986. Fisheries education and training: chasllenges and opportunities. Naga, The ICLARM Quarterly 9(4). 44 p. ICLARM. 1988. Research for the development of tiopial aquaculture: a proposed new v:nture for the Consultative Group on International

Agricultural Research (CGIAR). International Center for Living Aquatic Resources Management, Manila, Philippines. 40 p.LCA. 1980. ILCA. The first years. International Livestock Centre for Africa. Addis Ababa,Ethiopia. 127 p.

IRRI. 1974. An agro-climatic classification for evaluation of cropping systems in Southeast Asian rice growing regions. International Rice Research Institute, Los Bafios, Laguna, Philippines.

IRRI. 1982. Report of a workshop on cropping systems in Asia. International Rice Research Institute, Los Bafios, Laguna, Philippines. 756 p.Javier, E.Q. 1978. Integration of fodder production with intensive cropping systems in Southeast Asia, p. 19-36. In Feeding stuffs for livestock in

South East Asia. Malaysian Society of Animal Production, Serdang, Selangor, Malaysia. Johnson, S.11. and J.B. Claar. 1986. FSR/E: Shifting the intersection between research and extension. Agric. Admin. Ext. 21: 81-93. Lightfoot, C. 1987. Indigenous research and on-farm trials. Agric. Admin. Ext. 24: 79-89. Lin, It.R. 1982. Polycultural systems of freshwater fish in China. Can. J. Fish. Aquat. Sci. 39(1): 143-150. Mackie, A.M. 1982. Chemical stimulation of feeding behaviour, p. 275-291. In T.J. Ilara (ed.) Chemoreception in fisheries. Developments in

Aquaculture and Fisheries Science 8. Elsevier Science Publishers, Amsterdam, The Netherlands. Maclean, J.L and L.B. Dizon, Editors. 1987. ICLARM Report 1986. International Center for Living Aquatic Resources Management, Manila,

Philippines. 158 p. Mclvor, J.G. and C.P. Chen. 1986. Tropical grasses: their domestication and role in animal feeding systems, p. 55-60. In G.J. Blair, D.A. Ivory

and T.R. Evans (eds.) Forages in Southeast Asian and South Pacific agriculture. ACIAR Proceedings No. 12. Australian Centre for Intemraional Agricultural Research, Canberra, Australia.

Milstein, A., G. Ilulata and G. Wohlfarth. Multivariate analysis of relationships between management inputs and fish growth and yields in polyculture. (In press)

Mone D.C. 1978. Climate, soils and vegetation. 3rd ed. University Tutorial Press Ltd., Slough, UK. 286 p. Meriarty, D.J.W. and R.S.V. Pullin, Editors. 1987. Detritus and microbial ecology in ,iaculture. ICLARM Conference Proceedings Nc. 14,420

p. International Center for Living Aquatic Resources Management, Manila, Philippines. Munro, J.l. and D. Pauly. 1982. The ICLARM Network of Tropical Fisheries Scientists. ICLARM Newsletter 5(4):5. NACA. 1986. Report of the First Meeting of the Provisional Governing Council of the Network of Aquaculture Centres in Asia, 25-28 November

1986, Surabaya, Indonesia. NACA, Bangkok, Thailand. Naylor, E. and C. Scholtissek. 1988. Fish farming and aquaculture. Letters to Nature. Nature 333:506. Nugent, C. 1987. Small-scale systems for tilapia culture in tropical Africa - considerations for their widespread adoption. In R.S.V. Pullin, T.

Bhukasawan, K. Tonguthai and J.L Maclean (eds.) The Second International Symposium on Tilapia in Aquaculture. ICLARM Conference Proceedings 15. Department of Fisheries, Bangkok, Thailand, and International Center for Living Aquatic Resources Management, Maniia, Philippines.

Olh, J. 1986. Carp production in manured ponds, p. 295-303. In R. Billard and J. Marcel (eds.) Aquaculture of cyprinids. Institut Nationale de la Recherche Agronomique, Paris, France.

Oldeman, L.R. and M. Frdre. 1982. A study of the agroclimatology of the humid tropics of Southeast Asia. Tech. Rep. FAO, Rome. Opuszynski, K. 1981. Comparison of the usefulness of the silver carp and the big head carp as additional fish in carp ponds. Aquaculture 25: 223­

233. Pagt:io, M.C. 1987. Asian study tour for African and Latin American aquaculture researchers. Nags, The ICLARM Quarterly 10(3): 11. Pandian, T.J. and M.P. Marian. 1985. Nitrogen content of food as an index of absorption efficiency in fishes. Mar. Biol. 85: 301-311. Pauly, D. and K.D. Hopkins. 1983. A method for the analysis of pond growth experiments. ICLARM Newsletter 6(1): 10-12. Perkins, J., R.J. Petheram, R. Rachman and A. Semali. 1986. Introduction and management prospects for forages in Southeast Asia and the South

Pacific, p. 15-23. In G.J. Blair, D.A. Ivory and T.R. Evans (eds.) Forages in Southeast Asian and South Pacific agriculture. ACIAR Proceedings No. 12. Australian Centre for International Agricultural Research, Canberra, Australia.

Phiri, C.D. 1986. Adjusting the institutional environment of risk and uncertainty: implications for agricultural policy in Malawi. Agric. Admin. Ext. 23: 45-49.

Plucknett, D.L. and N.J.H. Smith. 1984. Networking in international agricultural research. Science 225: 989-993. Plucknett, D.L. and N.J.H.Smith. 1986. International prospects for cooperation in crop research. Econ. Bot. 40(3): 298-309. Popper, K.R. 1959. The logic of scientific discovery. Hutchinson, London, U.K. Prein, M. 1985 The influence of environment factors on fish production in tropical ponds investigated with multiple regression and path analysis.

Institut fOrMcreskunde an der Christian-Albrecht-Universitat zu Kiel. 91 p. Diplomarbeit. Pullin, R.S.V. 1985. Tilapias: "everyman's fish". Biologist 35(2): 84-88. Puffin, R.S.V. 1986. The worldwide status of carp culture, p. 21-24. In R. Billard and J. Marcel (eds.) Aquaculture of cyprinids. Institut de la

Recherche Agronomique, Paris, France.

51 Pullin, R.S.V. 1987. General discussion on detritus and microbial ecology in aquaculture, p. 368-381. In DJ.W. Moriarty and R.S.V. Pullin (eds.)Detritus and microbial ecology in aquaculture. ICLARM Conference Proceedings No. 14.420 p. International Center for LivingAquatic Resources Management, Manila, Philippines.Pullin, R.S.V. and R.A. Neal. 1984. Tropical aquaculhure. Need fora strong research base. Mar. Policy 8(3): 217-228.Pullin, R.S.V. and M.C. Paguio. 1987. The ICLARM Network of Tropical Aquaculture Scientists. Naga, The ICLARM Quarterly 10(3): 9.Pullin, R.S.V., T. Bhukasawan, K. Tonguthai and J.L. Maclean, Editors. The Second International Symposium on Tilapia in Aquaculture.ICLARM Conference Proceedings 15. Department of Fisheries, Bangkok, Thailand, and International Center for Living AquaticResources Management, Manila, Philippines.Rain, N.M., 0. Zur and Y. Avnimelech. 1982. Microbial changes occurring at the sediment-water interface in an intensively stocked and fed fishpond. Aquaculture 27: 63-72.

Reich, K. 1975. Multispecies fish culture (polyculture) in Israel. Bamidgeh 27(4): 85-99.Scholtissek, C. and E. Naylor. 1988. Fish farming and influenza pondemics. Nature 331:215.Schroeder, G.L. 1980. Fish farming in manure-loaded ponds, p. 73-86. In R.S.V. Pullin and Z.H. Shehadeh (eds.) Integrated agriculture­aquaculture fanning systems. ICLARM Conference Proceedings 4, 258 p. International Center for Living Aquatic ResourcesManagement, Manila, and the Southeast Asian Regional Center for Graduate Study and Research in Agriculture, College, Los Bafios,Laguna, Philippines.Schultze-Kraft, R. 1986. Exotic and native legumes for forage production in Southea.-t Asia, p. 36-42. In G.J. Blair, D.A. Ivory and T.R. Evans(eds.) Asian and Souts Pacific agriculture. ACIAR Proceedings No. 12. Austraiian Centre for International Agricultural Research,Canberra, Australia.Spedding, C.R.W. 1979. An introduction to agricultural systen.s. Applied Science Publishers, London, UK.Svirez.hev, Yu M., V.P. Krysanova and A.A. Voinov. 1984. Mathematical modelling of a fish pond ecosystem. Ecol. Modelling 21: 315-337.

icTemprosa, R.M., M.C. Paguio, M. Prein, R.S.V. Pullin and P. Edwards. Integrated .gi llture-aqu culture: a general bibliography. ICLARMBibliographies 4. (In preparation)Terra, G.J.A. 1954. Mixed-garden horiculture in Java. Malayan J. Trop. Geogr. 3: 33-43.Terra, G.J.A. 1958. Farm systems in Southeast Asia. Neth. J. Agric. Sci. 6(3): 157-182.Tripathi, S.D. and M. Ranadhir. 1982. An economic analysis of composite fish culture in India, p. 90-96. In Aquaculture economics research inAsia. IDRC-193e. International Development Research Centre, Ottawa, Canada.Van Dcr Lingen, M.I. 1959. Some preliminary remarks on stocking rate and production of 'alapia species at the Fisheries Research Centre, p. 54­62. Proceedings of the First Fisheries Day in Southern Rhodesia, August 19f.i Government Printer, Salisbury, Rhodesia.von Bertalanffy, L. 1968. General systems theory. Braziller, New York, USA.Venna, R.D. 1986. Environmental impacts of irrigation projects. J. Irrigation Drainage EngineeringWebster, C.C. and P.N.Wilson. 112: 322-330.1966. Agriculture in the tropics. The English Language Book Society and Longman, Lndon, UK. 488 p.Whittlesey, 1). 1936. Major agricultural regions of tlie earth. Assoc. Amer. Geogr. 26: 199-204.WIIO. 1980. Epidemiology and c"ntrol of schistosomiasis. WHO Tech. Rep. Ser. No. 643, 63 p.Wohlfarth, G.W. and G. llulata. 1987. Use of manures 3 53 3 6 7in aquaculture, p. ­ . InD.J.W. Moriarty and R.S.V. P,llin (eds.) Detritus andmicrobial ecology in aquaculture. ICLARM Conference Proceedings No. 14, 420 p. International Center for Living Aquatic ResourcesManagement, Manila, Philippines.Yashouv, A. 1971. Interaction between the common carp (Cyprinus carpio)and .hesilver carp (Ilypophthalmichthysmolitrix) in fish ponds.

Bamidgeh 23(3): 85-92.

APPENDIX I

WORKSHOP - TOWARDS A RESEARCH FRAMEWORK FOR TROPICAL INTEGRATED AGRICULTURE-AQUACULTURE FARMING SYSTEMS

15-17 October 1986, Manila, Philippines

A workshop was convened by ICLARM as part of the information-gatheriug process from which this study was produced. The workshop was supported entirely by UNDP and ICLARM and was held at ICLARM Headquarters, Manila.

The following papers were presented. Copies of the papers and a summary report may be obtained from Dr. R.S.V. Pullin, ICLARM.

International Research Cooperation in Wastefed Aquaculture and Integrated Farming Dr. Roger S.V. Pullin

Reseuxch Methodologies for the Development of Ti:pical Integrated Farming and Wastefed AquzCL ture Systems Dr. Kevin D. Hopkins

A Faiming Systems Research Approach to Integrated Agriculture-Aquaculture Dr. Joseph A. Gartner

Social Science and Economics Research Needs for the Development of Wastefed Aquaculture and Integrated Farming Dr. Ian R. Smith

Research for Training Needs for Wastefed Aquaculture and Integrated Farming Dr. Peter Edwards

The Research-Development Interface in Integrated Farming with Special Reference to Africa Dr. M.N. Kutty

Information Flow and Extension in Integrated Farming Systems Research and Development Mr. Jay L. Maclean

Research Priorities for the Development of Rural Aquaculture in Africa Dr. John D. Balarin

The Current Status and Future Potential of Tropical Integrated Farming Systems in Asia Dr. V.R.P. Sinha

Sociocultural Aspects of Integrated Farming

Technology Transfer from Asia to Africa Dr. Kenneth Ruddle

52

53

The participants were:

Mr. John D. Balarin Mr. Jay L. MacleanICLARM Project Leader, Africa Directorc/o Department of Fisheries Information ProgramPO Box 593 ICLARMLilongwe, Mala¢,i MC PO Box 1501

Dr. Barry A. Costa-Pierce Makadi, Metro Manila, Philippines Resident Consultant Scientist Miss Mary Ann C. PaguioPadjadjaran University Program AssistantJalan Sekeloa, Bandung Aquaculture ProgramWest Java, Indonesia ICLARM

MC PO Box 1501Dr. Peter Edwards Makati, Metro Manila, PhilippinesProfessor of AquacultureDivision of Agricultural and Food Dr. Roger S.V. Pullin

Engineering DirectorAsian Institute of Technology (AIT) Aquaculture ProgramGPO Box 2754 ICLARMBangkok 10501, Thailand MC PO Box 1501

Makati, Metro Manila, PhilippinesDr. Joseph A. Gartner Associate Professor of Farming Systems Dr. Kenneth RuddleDivision of Agricultural and Food National Museum of Ethnology

Engineering Seni Expo ParkAsian Institute of Technology (AIT) Suita, Osaka 565, JapanGPO Box 2754Bangkok 10501, Thailand Dr. V.R.P. Sinha

Senior AquaculturistDr. Kevin D. Hopkins Network of Aquaculture Centres in AsiaExecutive Director c/o National Inland Fisheries InstituteConsortium for International Fisheries Kasetsart University Campusand Aquaculture Development (CIFAD) Bangkhen, Bangkok 10900, ThailandOregon State UniversityCorvallis, Oregon 97331, USA Dr. Ian R. Smith

Director GeneralDr. M.N. Kutty ICLARMTeam Leader MC PO Box 1501FAO/African Regional Aquaculture Makati, Metro Manila, Philippines

Center (ARAC)PMB 6165 Aluu, Port Harcourt River State, Nigeria